You are on page 1of 196

The Game Behind the Video Game

Business, Regulation, and Society in the Gaming Industry

April 8-9, 2011 at The Heldrich in New Brunswick, NJ

Conference Proceedings

http://bit.ly/aKHOmu

Table of Contents
Mapping Gold Farming Back to Offline Clandestine Organizations: Methodological, Theoretical, and Ethical Challenges Brian Keegan, Muhammad Ahmad, Dmitri Williams, Jaideep Srivastava, Noshir Contractor Northwestern University, USA Trends In the Video Games Software Industry. Toward New Business Models? Giuditta de Prato & Jean Paul Simon European Commission, Spain Claudio Feijoo Universidad Politecnica de Madrid, Spain Hacktivism in Online Games: Negotiating Transparency, Privacy, & and Policy Making Peter Ludlow Northwestern University, USA Burcu S. Bakioglu Indiana University, USA Layers of Influence In Multiplayer Gaming Nadav Lipkin Rutgers University, USA A Business History of Video Games: Revenue Models from 1980 to Today Joost van Dreunen Columbia Institute for Tele-Information, USA The Active Video Game Consumer: What Is at Stake in the Narratives Surrounding the Video Game Prosumer Paolo Ruffino Goldsmiths, University of London, UK Call of Duties: The Arbitration of Online Game Disputes Ren Reynolds the Virtual Policy Network, London, UK. Melissa de Zwart The University of Adelaide, Australia Games are Not Coffee Mugs: Games and the Right of Publicity William K. Ford & Raizel Liebler The John Marshall Law School, USA New Media Censorship & the First Amendment: Understanding the Origins of EMA v. Schwarzenegger Thomas H. Rousse Northwestern University, USA Designing for Diversity: Can Regulation Promote Diverse Experiences in Game Worlds? Melissa de Zwart University of Adelaide, Australia Kate Roth Monash University, Australia Productivity Games: Improving Software Quality through Fun and Play Ross Smith Microsoft Corporation, USA Lets Play Super Rutgers RPG: Interactivity by Proxy in an Online Gaming Culture Kris Ligman University of Southern California, USA For Constitutional Control of New Media: Examining the Judicial Obsession with Causality in Video Software Dealers Association v. Schwarzenegger Ju Young Lee Penn State University, USA The Private Regulation of Virtual Worlds Versus Real World Laws: European and American Perspectives Sevan J. Antreasyan University of Geneva, Switzerland Copyrights Middle Ground: The Interaction of Players and Video Games Bruce E. Boyden Marquette University Law School

Mapping Gold Farming Back to Offline Clandestine Organizations: Methodological, Theoretical, and Ethical Challenges

Brian Keegan, Muhammad Ahmad, Dmitri Williams, Jaideep Srivastava, Noshir Contractor

Abstract Clandestine, covert, dark and illicit organizations are primarily characterized by the need to engage in coordination and collective action while also emphasizing secrecy and security (Ayling, 2009). However, empirical analyses of offline clandestine organizations structures have received scant attention because traditional data collection is difficult by design. Studies of clandestine organizations employ methods which censor their embeddedness within particular historical contexts and larger licit spheres of peripheral and legitimate actors. These studies rely on descriptive, single-level methods. However, the explosion of behavioral data available in online databases has opened up new avenues of social research. To the extent that individuals in online worlds operate under similar social and psychological motivations and constraints as the offline world, it is possible to use generative models of clandestine networks from online virtual worlds to test and inform theories of clandestine networks in offline contexts (Williams, 2010). We use gold farmers in massively multiplayer online games (MMOGs) as a case to examine how clandestine organizations assemble and maintain their operations in the face of pressure to remain competitive and secret. We review our recent research findings employing methods in network analysis and machine learning to detect and identify gold farmers in a popular MMOG based on distinct structural motifs in trade exchanges, patterns of behavioral similarity, and appropriation of in-game affordances. Although these findings on virtual clandestine organizations comport with many existing theoretical predictions as well as observations from offline criminal behavior, we discuss how they fail to map from online to offline in other contexts. Finally, we discuss the ethical implications of attempting to develop abstract

heuristics for identifying clandestine behavior in data rich contexts and conclude by identifying future directions for analytic and theoretical research.

Introduction The analysis of criminal and clandestine organizations presents unique theoretical and methodological problems. However, empirical analyses of these organizations have received scant scholarly attention despite pervasive popular ideas about terrorists, hackers, pranksters, and other malfeasors as flexible and evasive network organizational forms. This is primarily a consequence of the fact that the data on these organizations is necessarily had to come by; members obviously resist traditional data collection methods as they seek to avoid detection. Furthermore, the few analyses that have been done have been handicapped by a theoretical and methodological approach that emphasizes descriptive analysis of structural features at a single level of analysis but fail to account for how network topology and actor attributes influence each other. Recent developments in the statistical analysis of network data now permit confirmatory and inferential techniques, multilevel, multitheoretical analyses, and hybrid models that account for both network topology and actor attributes (Contractor, Wasserman, & Faust, 2006). By using these generative statistical approaches to network analysis, it is possible to compare networks despite differing widely in size, relationship, constituents, and time. The differences or similarities in these generative models allow us to analyze the extent to which task and other contextual features influence both micro-level actor tendencies and macro-level structural characteristics of clandestine social organization. In of themselves, these comparative network analysis methods do not solve the data scarcity problem outlined above. However, a comparative approach allows us to leverage observations from one domain to support inferences in another. In particular, the exhaustive behavioral data of individuals interactions and attributes in virtual spaces such as massively multiplayer online games (MMOGs) provides a venue to examine a variety of social and organizational processes, including clandestine and criminal activity. To the extent that individuals operate under similar motivations and constraints as the real world, it is possible to use clandestine data from virtual worlds to inform the development of theories of clandestine behavior in the real world.

A specific case of deviant behavior in online communities, gold farming, is first introduced to ground the review of social processes in covert network organizations. The possibilities and drawbacks for mapping online behavior to offline contexts are then discussed followed by a multilevel and comparative network analysis and review of the covert and criminal network analysis literature. The paper concludes with a discussion on the opportunities and limitations on using virtual world data to inform the development of computational social science in the domain of criminology, analyze, and map the social organization of illicit activities from the online world back to the offline world. Throughout the paper we will employ covert, clandestine, and illicit interchangeably to describe organizations primarily characterized by the need to engage in collaboration, cooperation, and collective action while also emphasizing invisibility, secrecy, control, and security (Baker and Faulkner 1993; Morselli et al. 2007).

Shortcomings of previous covert network analysis Traditional sociological and criminological examinations of deviant behavior have wholly avoided the relational and interactional forms of criminal activity (Vaughan, 2002; Waring, 2002). However, organizational theorists, communication scholars, and sociologists increasingly rely on network analysis methods to understand how actors within organizations interact and how these structures constrain and enable social behavior (Brass, Galaskiewicz, Greve, & Tsai, 2004) and only recently have network analysis methods been applied to understand how actors in dark networks, such as drug traffickers and terrorist cells, coordinate their activities and adapt their structure to achieve their mission while avoiding detection and maintaining resilience (Ayling, 2009; Milward & Raab, 2006). The organizations actors co-construct to support covert or illicit behavior are the primary focus of this paper as opposed to individuals covert/illicit behavior against overt/licit organizations (Pinto, Leana, & Pil, 2008). To this end, we adopt the perspective of organizational network theory where networks are not only the means of transmitting resources and information among a group of individuals, but are also the skeleton of upon which tacit and implicit practices such as hierarchies, status, and norms are built

they are sources of structure as well as carriers of process (Borgatti & Foster, 2003; G Robins, 2009). The organization forms assumed by covert networks and other secret groups demand conformity to extreme demands and thus serve as boundary conditions to test extant theories of organizing (Simmel, 1906). Rather than focusing on case or culturally specific constructions of organizational relationships, a structural approach examining more generalizable patterns of relationships predicated on transactions, communication, kinship, and authority permits comparative analysis into how individual agency and variation interact with structural tendencies and other exogenous features governing how covert organizations assemble themselves (Coles, 2001; Erickson, 1981). As with any organization, covert organizations assume a variety of forms in response to the shared goals, task demands, environmental factors, and skills of its members (Cressey, 1971; McIntosh, 1975). However, organizational studies of covert groups have traditionally been limited by the scarcity of data on these groups. Covert actors are usually a tiny minority of the general population, purposefully difficult to identify, and obviously resist entre from outsiders who wish them to disclose their relationships with trusted affiliates. These factors all complicate traditional data collection approaches such as observation or surveying. Because of these data collection difficulties, the theoretical and methodological literature examining covert networks has traditionally relied upon historical case studies, anecdotal accounts, or data gleaned from judicial processes (Davis, 1981; Sparrow, 1991). However, these approaches run afoul of network analysis research design. The choice of sampling units, the form of relations, the relational content, and the level of data analysis all bear on the validity of any network analysis (Knoke & Kuklinski, 1982; S Wasserman & K Faust, 1994). Problems of boundary specification abound in characterizing covert organizations as legitimate actors (e.g., lawyers and accountants who tacitly abet illicit activities) and peripheral facilitators (e.g., family members, corrupt officials) are simultaneously loosely-affiliated with the core tasks of the organization but nevertheless play crucial roles in enabling criminal activities (Morselli & Giguere, 2006). These actors are likely unobserved by investigators and therefore unrecorded in judicial proceedings or omitted from historical accounts (Sparrow, 1991).

Some overt organizations are assembled on an ad-hoc basis, improvise, and have substantial turnover in membership, while others are highly stable, routinized, and dependent on trust, and both of these (McIntosh, 1975). The nature of relationships between actors in a network analysis likewise differentially affects validity of any analysis; criminal organizations depend upon a variety of transactional, communicational, instrumental, kinship, and authority relations which independently only capture a single dimension of the network and certainly do not reliably align themselves in every case (Coles, 2001). These relations are not binary either, the strength or weakness of ties in these organizations are proxies for durability, intensity, duration, frequency, and trust. Approaching these relations as dichotomous entities has substantial topological implications for the density, reachability, centrality, and equivalence of the network. Finally, although network analysis permits the collection of data at multiple levels of analysis (actors, dyads, triads, groups, organizations, etc.), many types of network data unfortunately are transformed and simplified into a single level of analysis under simplifying assumptions. The literature is rife with examples of dyad atomization (M. Granovetter & Swedberg, 1992) which ignore the possibility that the larger structural context may also play a role (Jones, Hesterly, & Borgatti, 1997). The graph theoretic approaches very much in vogue among physicists, mathematicians, biologists, and other domains leading the new science of networks in particular emphasize topological features and distributions of graphs such as degree distribution, geodesic lengths, and other descriptive characteristics (Newman, 2003). In the context of interactions among populations of homogenous entities such as atomic particles, proteins, or routers, the assumption that qualities of the actors have little or no variance or should not influence linking patterns may be justifiable. However, this claim is completely untenable for human social systems, especially studies of covert and clandestine organizations where variation among individuals attitudes, predispositions, motivations, and resources fundamentally interact with the processes that cause them to have relationships. The topological approaches employed by many physicists, biologists, and mathematicians fail to model the variance among individual actor attributes and how these in turn influence local and

global network structures. Because covert behavior requires individual action, covertness is necessarily at the level of the actor and cannot be the construct of the network itself (Robins 2009).

Multi-level processes of covert organizational assembly Covert organizations rely upon processes of social exchange and reciprocity regulate the provision and distribution of illicit goods, protections, regulation, and extortion of these services (McIllwain, 1999). In the context of clandestine networks where balancing the twin purposes of covertness and collaboration are paramount (Baker & Faulkner, 1993; Morselli, Gigure, & Petit, 2007), trust, in particular, is the foundational relationship for subsequent types of interaction such as communication, exchange, and dominance (von Lampe and Johansen 2004). Individualized trust proceeds from personal qualities or past behavior, reputational trust is developed on the basis of others relations/endorsements, and shared trust based on group membership. Processes of social selection are also likely to dominate the formation, maintenance, and dissolution of these trust-oriented antecedent relationships as individuals fixate on shared backgrounds and values; the more these overlap, the stronger the tie is likely to be and the greater the likelihood that the dyad interacts in other relational contexts as well (Erickson, 1981; von Lampe & Johansen, 2003). Other individual-level attributes such as trust predicated on identification via familial ties, shared background, role specialization, or comparative advantage may also govern organizational assembly practices (Coles, 2001). Similarly, balance theory processes such as common adversaries (i.e., al Qaida and the Taliban both resist the U.S.) and exchange theory processes such as resource dependence (arms dealers supplying both sides of a conflict) are also examples of structural factors governing the creation of strong trust ties (G Robins, 2009). Following Granovetter (1973), the existence of these strong ties should lead to the accumulation of closed, clique-like structures. While these structures have advantages by allowing members to monitor and sanction each others behavior to ensure control, loyalty, and covertness, the presence of many strong ties can increase the risk of detection. It may be the case that cognitive-cultural mechanisms such as

shared background and ideology can also internalize organizational imperatives for control, loyalty, and covertness to foster trust in the absence strong relational ties of communication or exchange (Milward & Raab, 2006). These mechanisms can support trust within particular cliques and cells as well as maintaining identification between cells despite the lack of strong ties and possibility that inter-clique ties can be revealed and disrupted (G Robins, 2009) However strong ideological identification and commitment is not universal to all types of criminal organization (Cressey, 1971; McIntosh, 1975), and the theoretical paradigms outlined above still exhibit a tendency towards densely connected cliques/cells rather than decentralized, but still wellconnected, covert organizations. In particular, brokers are highly autonomous agents who bridge structural holes between the cliques or core of the network to the periphery of the network can serve multiple purposes. These may be actors strategically employed by the leadership to insulate themselves from outside interdiction (Dorn, Oette, & White, 1998), entrepreneurs brokering information, or legitimate actors facilitating illicit activity (Morselli & Giguere, 2006). Brokers are vulnerable insofar as both sides of a structural hole can betray them. However, the loss of these brokers can also significantly disrupt the effectiveness of the network (Morselli, 2010). The fact that brokers remain can autonomous reflects the fact that they much less likely to buy into the ideological identification or share the identity that motivates core participants. Thus, it is difficult to identify brokers by extrapolating the behavior of typical network members to these individuals in this distinctive roles (G Robins, 2009). Actors in covert organizations likewise face a major dilemma between ensuring operational secrecy and efficiency (Ayling, 2009; Milward & Raab, 2006). Networks characterized by high centralization and low diameter are associated with high levels of performance (Bavelas, 1950; Guetzkow & Simon, 1955), although these findings are primarily based on small groups in a laboratory setting with specific types of tasks. More recent analyses have emphasized that teams with dense ties as well as leaders who are central in intergroup networks are more committed and successful (Balkundi & Harrison, 2006). However, traditional bright organizations operate under considerably different constraints than dark organizations like gangs, drug traffickers, terrorists, or arms dealers where dense networks of ties and prominent bridging leaders are

liabilities which increase the likelihood of detection (Ayling, 2009; Baker & Faulkner, 1993; Raab & Milward, 2003). Clandestine operations must to balance efficiency with security (Morselli, et al., 2007) and organizational resilience with operational flexibility (Milward & Raab, 2006). How do covert organizations balance security with efficiency then? Ericksons study of six diverse clandestine organizations concludes organizations with an established reputation are committed to emphasizing security over efficiency (Erickson, 1981). Baker and Faulkners study of price fixing and collusion in a white collar crime ring revealed that peripheral players were less targeted and less sanctioned than more central players (Baker & Faulkner, 1993). Decentralization has also been observed to be a key tactic adopted by members of a criminal network in response to targeting and asset seizure by lawenforcement (Morselli & Petit, 2007). Compared to the network of al-Qaida terrorists, a drug trafficking enterprise engaging in regular activity exhibits higher centralization and a core of closely-linked participants with stable roles (Morselli, et al., 2007). The addition of other actors to the core of a criminal network can serve to extend its periphery and insulate participants at the core (Dorn, et al., 1998). Computational models attempting to optimize efficiency against likelihood of detection will still assume star-like structures such as wheels and windmills (Lindelauf, Borm, & Hamers, 2009). However, these are all distinct and potentially dynamic processes and this research has only examined them at a single level of analysis; taken together, how do the tendencies towards or away from reciprocity, closure, preferential attachment, and homophily simultaneously influence the ultimate structure of the network? A multi-theoretical, multi-level model of clandestine activity thus needs to account for a variety of simultaneous tendencies outlined above: some covert actors will seek balance and closure, other entrepreneurial actors will have tendency to bridge structural holes and remain unembedded, efficiency drives tendencies towards centralization, but exogenous threats of detection demand low densities.

Toward a multi-theoretical, multi-level framework

Although the novelty of contrasting emergent and formal organizations has waned and been replaced by an emphasis on examining how organizational structures emerge (Contractor, Wasserman, Faust 2006), few analyses of clandestine networks have been performed let alone modeling the dynamics by which organizational forms emerge. As Robins (2009) argued, How [does] the network provide constraints and opportunities for the individuals within it and how are those individuals [attempting] to shape the network to maximize the opportunities and limit the constraints. Do the individual-level and network-based effects have independent explanatory capacity, do they interact in important ways, or can one be subsumed parsimoniously by the other? Monge and Contractor (2003) identify nine broad classes of theoretical mechanisms that have been used to explain the creation, maintenance, dissolution, and reconstitution of organizational networks. Crucially, because these theoretical mechanisms operate at multiple levels of analysis, they can potentially lead to confounding, complimentary, or contradictory manifestations in network structure. It is important to specify models that control for the respective contributions of endogenous and exogenous mechanisms that can also potentially account for what network ties are created, maintained, and dissolved (Contractor, et al., 2006; Monge & Contractor, 2003) There are at least five levels of analysis one might employ for network analysis (Contractor & Monge 2003, Contractor et al. 2006, Robins 2009). we review these below and identify the shortcomings of only examining each particular level of analysis. Across all these levels of analysis, parallel processes of selection and influence occur. Selection processes emphasize how individuals select certain partners or positions in the network based on individual attributes while influence processes emphasize how individuals attributes are influenced by their position in the network and behavior of their partners. Individual level factors. we previously outlined the shortcomings of topological approaches to network analysis for often failing to account for variation in individual-level factors. Certainly, some factors such as demographics (gender, ethnicity, geographic origin, family) are fixed within individuals but vary substantially between actors. In a network composed of friendship or sexual relations, one would

10

clearly expect an individual-level factor like gender to strongly influence the evolution of either type of network. Criminological studies emphasize how geographic origin or familial relations are the foundational attributes by which criminals evaluate whether or not to trust an individual (Erickson, 1981; von Lampe & Johansen, 2003). On the other hand, individual-level capacities such as age, expertise, or knowledge (e.g., knowing who is an informant) status might be expected to vary over time. In the context of criminal network analysis, individuals capacities to learn new information may strongly influence who they trust and thus with whom they interact. However, individuals information states are not independent, but are can be altered as a result of other actors beliefs which can diffuse like a disease via shared relationships in the network. Individuals psychological and cognitive factors are likewise internally held and dynamic states which may predispose them to be risk-takers, charismatic, subservient, or violent. These psychocognitive factors clearly influence processes of social organization in clandestine and criminal contexts by governing the roles assumed by various actors. Simply looking at individuals psychological states absent network structure potentially omits how positive or negative affect can develop as a result or the presence or absence of ties with other parties. Balance theory, for example, emphasizes how the absence of a friendship or trust tie between two individuals with a mutual friend or trusted counterparty leads to heightened stress and less affect (XXX). Finally, individuals may have various types of economic capital which can be lost or exchanged as a result of interactions with members of their own organization, competing organizations, or law enforcement. Dyadic level factors. A dyad is a pair of individuals and the relationship(s) they do or do not share. A pair of actors may have multiplex ties, in which they share many types of relationships such as trust and communication. The tie may also be directed (As relationship with B is distinct from Bs relationship with A), reciprocated (mutuality or exchange may cause A and B to have ties with each other), valenced (positive or negative, dominant or submissive), or valued (strong to weak, short-term to long-standing) (Robins 2009). As before, if this is the sole level of analysis, it omits important

11

interactions with other levels of analysis. For example, the likelihood of a dyad having a relationship may be a function of their mutual similarity (homophily) or dissimilarity (heterophily). The presence or absence of a tie could also be affected by exogenous factors such as the number of other ties each of the actors have (adding ties above ones Dunbar number may be very costly or unlikely). Local structural factors. As has been alluded to before, the presence of absence of ties in a network may be a function of the structure of existing ties irrespective of individual attributes. Processes of closure are particularly observed in trust-based relationships such as friendship where individuals with mutual friends themselves become friends with each other. Depending on the context, other structural processes such as transitivity or cyclicality may influence the types of ties created. A variety of definitions also exist to specify individuals membership and embeddedness in sub-groups defined by a shared pattern of ties with each other such as cliques, clans, cores, and plexes (Hanneman & Riddle 2005). A clique is a sub-group in which every member of the clique has a connection to every other member of the clique. Unfortunately, analyzing these meso-level structures alone confounds unique processes that can generate similar structures; reciprocation in a triadic context is qualitatively different from reciprocation in a dyadic context (Krackhardt, 1999). Glocal factors. Individual actors can have properties that are functions of the rest of the actors position and behavior in the network. Traditional approaches to network analysis have emphasized understanding descriptive statistics such as centrality. These metrics capture how many connections a node has (degree centrality), how proximate it is to the rest of the nodes in the network (closeness centrality), the extent to which it spans structural holes in the network (betweenness centrality), or is connected to other well-connected nodes (Bonacich power or eigenvector centrality) (Wasserman & Faust 1994). Although they begin to bridge local and global levels of analysis, centrality metrics capture superficial structural tendencies and can confound individuals with very different attributes and/or qualitatively different positions and roles. Global factors. It is also possible to define properties using the whole network as a unit of analysis. Measures such as diameter describe the length between the two most distant, but indirectly

12

connected nodes; density capture the prevalence of ties relative to the theoretical maximum number of ties; centralization conveys the distribution of links among nodes; and assortativity suggests whether well-connected nodes tend to have other well-connected nodes as neighbors (Newman, 2003).

Comparing generative, multi-level statistical models of networks McIllwain (1999) asserts that networks as social system of organized crime exhibit remarkable consistency of the process of organizing crime across time and space. This is an empirical claim, but testing this assertion has only become possible in recent years because in addition to the difficulty of obtaining data, the complex dependencies within network data and confounding processes that generate similar topological features makes it difficult to readily compare network structure. Unlike other types of data collected from individuals, relational data used in network analysis usually show strong interdependencies; the existence of a tie from A to B is not independent from the observation of other ties involving A and B. For example, the existence of one friendship relation may depend on the total number of friends the focal node has, what other friends the focal node has and/or the relations between the existing friends and the potential friends. The interdependencies among relational data make it inappropriate to analyze network data using traditional statistical analysis approaches, such as regression analysis or ANOVAs (S Wasserman & K Faust, 1994). As we review in greater depth in the methods section, statistical models like exponential random graph models (ERGMs) allow us to test hypotheses about the network without losing information about interdependencies among relations by estimating the probabilities that local network substructures are observed more frequently than by chance. In an exponential random graph model, the range of possible networks and their probability of occurrence under the model are represented by a probability distribution on this set of all possible graphs. The model is constructed by specifying parameters corresponding to the structural characteristics(s) which are more likely to occur in the distribution. There are numerous structural tendencies to characterize these networks and the particular parameters employed must be theoretically relevant such as reciprocity, transitivity, cyclicality, and star tendencies. These structural

13

tendencies, in turn, can be interpreted within a multi-theoretical framework that accounts for various endogenous and exogenous processes occurring at multiple levels of analysis that potentially govern the evolution of the graph (Contractor, et al., 2006; Monge & Contractor, 2003). The majority of network analyses are cross-sectional case studies of individual communities and do not attempt to compare networks. Two straightforward cases of comparative network analysis examine how the same set of actors either change their patterns of interactions over time or have different structures for different types of relationships. Studies of organizational networks reveal informal structures based on trust and friendship overlapped and revealed latent leaders (Kilduff & Krackhardt, 2008). A third type of analysis examines when the same relation is measured on two or more sets of actors. The Add Health studies measure friendship networks of adolescents across different schools permitting analysis of general properties of adolescent friendship formation (Moody, 2001). A fourth type of analysis asks whether data on similar types of relations from different actors in roughly similar settings and contexts are comparable. Holland and Leinhardt (1970) used triadic census approaches to examine the distribution of structural features across various network types. However, the most challenging type of comparison involves networks of different sets of actors of very different sizes and with substantially different relationships. Thus, to compare networks, we adopt the approach suggested by Faust and Skvoretz (2002) that networks are similarly structured to the extent that they exhibit the same structural tendencies, to the same degree. By characterizing each network using a statistical ERGM capturing the likelihood for the graph to assume structural properties, it will be possible to measure the similarity between networks based on these parameters estimates, represent the similarities among networks, and interpret the resulting distributions of tendencies using contextual information about the networks. Despite the fact that networks can vary substantially in size and density, these are differences of scale rather than process. Two networks with estimates similar in both magnitude and direction are similarly structured because the same structural tendencies are important and important to the same degree in predicting tie probabilities in both networks. It is possible to calculate pair-wise

14

(dis)similarity statistics for each network using Euclidean distances or correlation coefficients for predicted tie probabilities or direct parameter estimates themselves (Faust & Skvoretz, 2002).

Mapping behavior from the online to the offline To the limited extent that networks of criminal or clandestine have been analyzed, the data has generally been obtained from court documents such as wiretap records to capture whom talks to whom (Davis, 1981; Sparrow, 1991). However, this approach is limited for several reasons. First, it omits the embeddedness of these practices in larger social structures and the enabling roles played by legitimate and peripheral actors as discussed above. Second, it only captures particular types of relationships (such as phone communications) and the subset of actors that are parties to the relationship instead of than the multiplex ties that bind actors together. Third, these data are often cross-sectional and/or temporally censored; they fail to capture the communication dynamics preceding the onset of surveillance, how the organization was formed, or how these interaction patterns changed over time as individuals entered and left the system. Finally, the data often lack rich actor-level attributes such as demographics (gender, ethnicity, age), capacity (skill, expertise, knowledge), psychology (attitudes, motivations, beliefs), possessions, or behaviors. However, the explosion of behavioral data available in online databases has opened up new avenues of social research (Lazer et al., 2009). Massively multiplayer online games (MMOGs) are largescale social environments contain players of varying levels of expertise who join cooperative teams to accomplish complex tasks (Huang et al., 2009; Huffaker et al., 2009). To the extent that individuals in online worlds behave and interact under similar constraints and motivations as they do in an offline context, online worlds may provide a parallel in which researchers can test general hypotheses of human behavior (Bainbridge, 2007; Williams, 2010). Moreover, because the organizations that operate MMOGs maintain archival databases of all player actions and attributes, it is possible to analyze comprehensive cross-sectional and longitudinal behavioral data on a scale that would be unethical, impracticable, or

15

impossible to do in the real world. This is particularly exciting in the context of analyzing covert organizations. However, virtual worlds also offer substantially different affordances than the real world: death is impermanent, physics are negotiable, and consequently human behavior in them may not map at all from the online back to the offline. It is crucially important to establish three constituent components of mapping: face validity, concurrent validity, and predictive-external validity. Face validity captures whether the measure reflects the phenomenon of interest. In the context of understanding how clandestine groups organize themselves, it is necessary to establish that there are individuals who both act in a covert manner and interact with each other to accomplish shared goals. In the context of a game like EverQuest 2 where players interact and engage in behaviors known to lead to administrative banning, there appears to be face validity on the mapping online clandestine activities to offline clandestine activities as there are motivations and incentives to organize in both. Concurrent validity is the extent to which the measured phenomenon correlates to other measures of the same phenomena by checking against other criterion. In both contexts, members of clandestine organizations should feel the pressure to operate secretly and avoid detection, the general population should see these covert organizational members as deviants that authorities should control, and online administrators should allocate resources to detect and actions to destabilize these organizations. These remain open empirical questions for future research. Predictive-external validity tests whether a measure relates to other measures as expected. Not only must the measure predict the appropriate outcome in the virtual world, but the behavior must operate the same as one would expect in the real world (Williams, 2010). If the social organization of covert actors in an online game is to have external validity offline, the act of banning someone from the game should be similarly punitive as detainment, prosecution, and forfeiture of freedom and property in the offline context such that it motivates covert organization and promotes similarly evasive responses. However, the second standard of predictive validity by characterizing as operating as one would expect in

16

the real world is complicated by the discussion above that little is known about the particular dynamics of many clandestine operations there is nothing to compare it against. Clearly, there are very diverse forms of covert organizations as well as substantial variation in the code supporting the social architecture of virtual worlds. It is essential to identify a particular type of practice in a virtual world that has a ready and proximate mapping to offline virtual world practices. we suggest that gold farming, a practice involving accumulating and distributing virtual currency, bears a strong resemblance to the task demands of drug trafficking or money laundering. we review the background on the practice in the next section and describe some parallels it offers to support the mapping hypothesis.

Gold farming as a model type of online deviance Massively-multiplayer online games (MMOGs) such as World of Warcraft, EverQuest II, and Lord of the Rings Online are examples of fantasy-based game worlds in which millions of players interact in a persistent virtual environment. Just as these game economies exhibit macroeconomic characteristics observed in real-world economies (E Castronova et al., 2009), virtual worlds also contain black markets for acquiring goods and skills. Gold farming and real money trading refer to practices that involve the sale of virtual in-game resources for real-world money via exchanges outside of the game itself. The name stems from a variety of repetitive routines (farming) which are employed to accumulate virtual wealth (gold) which is sold to other players who lack the time or desire to accumulate their own in-game capital (J. Dibbell, 2006; Heeks, 2008). Gold buyers purchase this virtual capital to obtain more powerful weapons, armor, and abilities for their characters. This, in turn, accelerates players to higher levels and allows them to explore larger parts of the game world, confront more interesting and challenging enemies, and increase their social standing (E Castronova, 2005). Gold farming has been constructed as a deviant activity by both the game developers as well as the player communities for a variety of reasons. First, in-game economies are designed with carefully-calibrated activities and products that serve as sinks to remove money from circulation. Because gold farmers and

17

buyers inject currency into the economy, they create inflationary pressure, unintended arbitrage opportunities, and other perverse incentives for market agents. Second, farmers activities often overtly affect other players experiences by excluding them from shared game environments, employing anti-social computer scripts (bots) to automate the farming process, and engaging in the outright theft of account and financial information from their customers (Heeks, 2008; G. Lastowka, 2006). Third, the game developers are risk-averse to the legal implications (such as property rights, taxation, and torts) of sanctioning a multinational industry estimated to generate between $100 million and $1 billion in revenue annually (Edward Castronova, 2006; Lehtiniemi, 2007) while lacking legal jurisdiction, precedent, or regulation (J Dibbell, 2003; F. G. Lastowka & Hunter, 2006). Finally, farming upsets the meritocratic and fantasy-based nature of the game; some players may cease to play if other players can buy rather than earn accomplishments (Consalvo, 2007). For these reasons, game developers actively and publicly ban accounts engaged in gold farming (Tyren, 2006). Like other types of criminal organizations and dark networks, the pressure exerted by game developers to identify and disrupt gold farmers requires to balance tendencies towards decentralized networks which provide the greatest security and resilience against the high efficiency and trust of centralized networks (Morselli, et al., 2007). Because gold farming networks potentially operate under similar constraints as other covert organizations, the structure and dynamics of these organizations can be used to characterize and understand deviant and criminal activity such as drug trafficking or money laundering. Drug trafficking chains are characterized by non-hierarchical, non-vertically integrated supply chains of producers, distillers, distributors, mules, dealers, and users (Brzezinski, 2002; Raab & Milward, 2003). Both types of activity involve accumulation and distribution of commodities with large profit margins in exchange for cash. Because exchanges from individual to individual are highly dangerous are dangerous to the rest of the organization as the break of any one link can allow authorities to back-track and roll up the rest of the organization, these organizations distribution channels should be characterized by high levels of brokerage/structural holes (alternatively, a lack of closure) and large distances to insulate core members

18

from infiltration from the outside. Although organizations operate in a competitive marketplace, the demands of balancing secrecy and efficiency require the actors forgo exclusivity and allows individuals contract with individuals at various other levels. Thus, we might expect that gold farmers and drug traffickers are both characterized by large, but weakly connected giant components. Furthermore, because the organization is operating extra-judicially, disputes cannot be resolved by appealing to the authorities but must be negotiated and enforced in other ways. In this context, the mapping potentially crumbles; where narcotics organizations rely on ethnically homophilous distribution channels to ensure kinship-based trust as well as violence to enforce compliance, interactions and transactions among gold farmers are online manifestations of situated offline organizations that need not manifest itself with similarly homophilous selection biases along demographic categories of character race, class, or gender. However, it may be the case that homophily operates along other attributes such as play style/frequency or character level/age. We might also expect that gold farmers and drug traffickers could potentially rely upon unaffiliated legitimate actors who tacitly or unknowingly abet the operation (one might imagine a fellow dungeon raid member could be as unsuspecting of his or her partner as ones spouse being a dealer). These legitimate and peripheral actors may support covert organizations becoming larger and more efficient than if they remained perfectly insular. Moreover, as was the case with Baker & Faulkners analysis of a price fixing conspiracy, the most central actors in a gold farming or drug trafficking network are actually strategically positioned to increase their likelihood of detection relative to the true leaders who seek to avoid detection. As such, covert organizations may have heuristics about which behavioral patterns authorities are known to select on and have structured their operation such that these actors are more likely to be identified and banned while they can continue to rely on peripheral actors who remain undetected via other types of relationships or behavioral attributes. However, the mapping laid out here is imperfect. While traditional criminal organizations rely on trust to recruit co-offenders and engage in their tasks (von Lampe & Johansen, 2003), gold farmers can effectively create new or replacement co-conspirators at low costs by registering new accounts. Thus, gold

19

farmers, unlike other criminals, do not need to recruit or convert existing players, engage in processes of ideological identification or control over existing members, or threaten violence against defectors to accomplish their mission (Milward & Raab, 2006; G Robins, 2009). The consequences of being identified and banned are largely pecuniary rather than the long-term or even mortal risks in the real world.

Methods Statistical modeling of social networks Techniques for measuring local and global properties of a social network have existed for decades. Metrics such as centrality, clustering, density, diameter, cliques, and equivalence fulfill valuable roles for describing the properties of nodes, subgroups, and the whole network. Nevertheless, these approaches are only descriptive in nature. Just as computing means, distributions, and pair-wise significance tests do not themselves provide meaningful insight into the relationship between gender and income (for example), simply calculating the aforementioned descriptive network metrics does not permit inferences about the underlying social and behavioral processes that govern how social networks are created or maintained. A statistical approach to network analysis uses generative approaches to simulate similar network structures, evaluate the differences, and thus confirm whether particular structural processes influenced the configuration of the network more than by random chance. Furthermore, because network data is fundamentally relational, these data violate the fundamental assumptions of traditional statistical approaches such as linear or logistic regression. A regression model of whether or not a tie exists between a pair of actors would assume the presence of a tie is independent from the occurrence of other ties in the network and probability for each tie being created is identically distributed. In reality, this is an unreasonable assumption as the presence of a friendship tie between two actors in a social network may in fact be strongly influenced by a variety of exogenous factors such as whether the pair of actors have a friend in common, share the same race or gender, and the number of other friends they have.

20

Developing a statistical model of social networks thus requires a framework which accounts for the intrinsic interdependencies in network data. Over the past 25 years, a class of statistical models called P* or exponential random graph models (ERGM) have been developed which explicitly incorporate dependence assumptions for network analysis (Anderson, Wasserman, & Crouch, 1999; Frank, 1981; Frank & Strauss, 1986; Pattison & Wasserman, 1999; Garry Robins, Pattison, Kalish, & Lusher, 2007; G. Robins, Pattison, & Wasserman, 1999; S. Wasserman & K. Faust, 1994; Wasserman & Pattison, 1996; Wasserman & Robins, 2005). Following Robins, Pattison, Kalish, & Lusher (2007), there are at least five distinct motivations for developing statistical models of network data. 1. Social behavior is a highly complex system which implies that even small changes can dramatically change how the system evolves and stabilizes. Rather than trying to make perfect deterministic predictions, the models are stochastic and this noise allows us to understand the distribution of possible outcomes for a model and the associated uncertainty of its estimates. 2. Statistical models of networks require the specification of statistical parameters and the estimates for these parameters indicate whether or not the structure occurs more or less often than would be observed by chance. Thus, analogous to multivariate regression, a statistical modeling approach allows for hypothesis testing about multiple social processes that may lead to the over- or underrepresentation of structural configurations of theoretical interest. 3. The presence of ties, clusters, or other structures in the network may emerge as a result of distinct processes occurring in parallel. As such, simply counting frequencies of these structures potentially confounds a variety of processes which could have lead to their emergence. Again analogous to multivariate regression, a statistical model of networks should include controls which allow the analyst to evaluate the relative contribution of each main effect to the networks structure as well as the possibility that these independent parameter variables also interact with each other to generate particular network configurations. 4. An accurate and parsimonious model of the processes which govern the structure of a network can be a highly efficient representation of the network. This is substantively important because

21

accurate statistical models permit comparative network analysis by assessing the similarities or differences between the models themselves rather than either single-level descriptive statistics or large, complex complete representations. 5. Because global network structure emerges from local processes, statistical models of networks should rely upon parameters that are specified at a local level. Exponential Random Graph Models P*/ERG models make several assumptions. First, the number of nodes is fixed: models are compared against networks of the same size and do not account for the introduction or removal of nodes. Second, the models only predict the presence or absence of ties, rather than the value (i.e., strength or frequency) or valence (i.e., positive or negative affect) of a tie. Third, each ties in the network is regarded as a random variable. Fourth, the model assumes a self-organizing and stochastic process of generating ties based on the surrounding context (what ties already exist). Fifth, a model is composed of parameters which correspond to the structural features of interest and apply homogenously across the network. Taking a friendship network as an example, one might expect that processes of reciprocity would occur over and above the chance that reciprocated ties occur at random. A model to test this process would then consist of two parameters: a parameter corresponding to the tendency for links to occur at random and a parameter corresponding to the additional tendency for links to be reciprocated. If there is a reason to suspect that reciprocation should more often in a given part of the network (women are more likely to reciprocate than men, for example) rather than occurring uniformly across the network, then this process needs to be modeled with yet another parameter (such as a gender attribute effect). The observed network is only one possible realization of links among a given set of actors and the goal is propose a model for the stochastic process which generated this network. We can evaluate the extent to which a particular process contributed to a networks structure by comparing the observed network to other networks generated by the hypothesized process. A statistical model thus assigns a probability to all possible networks such that networks exhibiting properties of interest are more likely to

22

be used as a basis for comparison. In the friendship network example, if high levels of reciprocation are found in the observed network, the model should assign higher probabilities to the simulated graphs that also have high levels of reciprocation and use these as the basis for comparison. However, the sample space permitted by networks relational data is enormous because the addition or removal of a single link between any pair of actors necessarily results in a different network. These other networks may have substantively similar local and global metrics or very different structural properties. The whole sample space of a network includes the completely empty graph in which no connections are present between any nodes, the completely connected graph in which a connection is present between every pair of nodes, and a huge number of permutations in between. The number of permutations for a network is a function of the number of nodes (n) it contains. For a directed graph, there are possible permutations for the network while a directed graph has possible

permutations. Thus, a directed network of just 17 nodes has 2272 potential configurations which means there are more unique configurations for a 17-node network (2272 7.58x1081) than there are atoms in the universe (~1x1080). Because evaluating and assigning these probabilities to each distinct configuration of the network obviously becomes computationally intractable for graphs of even a seemingly trivial size like n=17, statistical models relying upon sampling and maximum-likelihood estimation are employed instead to generate parameter coefficients for p*/ERG models. The estimation of parameters relies upon a maximum likelihood criterion to choose parameter values such that only the parameter which maximizes the structural tendency found in the observed network is selected for the model. If the parameter for reciprocity is equal to zero, then reciprocation occurs purely by chance. If this parameter is positive, then reciprocation occurs more often than one would expect by chance. If the parameter is negative, then reciprocation occurs less often than chance. Using a Markov chain Monte Carlo maximum likelihood estimation (MCMCMLE) approach, a distribution of random graphs is simulated based on a starting set of parameter values, these values are refined by comparing the distribution of the simulated graphs against the observed graph, and the process

23

is iterated until the parameter estimates stabilize (Handcock, 2003; Snijders, 2002; Snijders, Pattison, Robins, & Handcock, 2006). However, stabilization and convergence of parameters is not guaranteed. On one hand, a proposed model may be insufficiently specified such that the observed network is unlikely to occur even under the maximum likelihood coefficients. For example, the proposed model may vastly over- or underestimate the amount of reciprocation or transitivity which is actually observed in the network. Using a goodness-of-fit approach to generate simulated networks and compare the resulting distribution of their network metrics (i.e., degree distribution, diameter, embeddedness, etc.) against the observed network is a necessary confirmatory step for evaluating the validity of an estimated model. On the other hand, estimation of models can also result in degeneracy when a model implies that every possible graph has a very low probability or the completely connected or unconnected graphs have a very high probability (Handcock, Hunter, Butts, Goodreau, & Morris, 2008; Hunter, Goodreau, & Handcock, 2008). Alternatively for large networks, a model may fail to converge because it maximizes on a set of parameter values without sufficiently exploring the sample space for other maxima or because it lacks parameters which account for higher-order non-Markov processes in large networks (S. M. Goodreau, 2007; Garry Robins, Snijders, Wang, Handcock, & Pattison, 2007). Similarly, overspecification of models with too many parameters results in networks which are substantively difficult to interpret, tremendously expensive to compute, and have statistically problematic estimates and standard errors given the loss of degrees-of-freedom. As a consequence, the estimation of networks using P*/exponential random graph model approaches is as much an art as it is a science as it requires the inclusion of theoretically well-motivated parameters as well as careful model construction, iteration, and evaluation of model fit. The modeling approach and methods used in this analysis are outlined below.

Gold farmer networks in EverQuest II Anonymized database dumps were collected from Sony Online Entertainments massivelymultiplayer online game EverQuest II. These data include both cross-sectional attribute data about

24

individual characters as well as longitudinal data cataloging character-to-character transactions. Because activity within the game is spread amongst several unique servers running instances of the game in parallel, a record of player-to-player exchanges on a single, representative player versus environment (PvE) server was condensed to generate a weighted, directed edge list of all transactions between characters on that server between for 36 weeks between January and September 2006. A separate table recording instances of accounts banned by the developer for abuse, non-payment, and other reasons was parsed to extract rationales related to plat, spam, farm, gold, coin, bot, and launder. The list of banned gold farmers was intersected with the trade activity database based on common account identification numbers to identify counterparty characters in which at least either the transaction sender or receiver had been banned by game administrators using gold farming-related rationales. Just as a list of judgments from criminal proceedings is not an exhaustive account of all criminal activity, the cancellation table is not a complete list of all gold farmers. Previous research using a machine learning approach to classify gold farmers based upon demographic and behavioral variables generated a large number of false positives, which may be evidence of unidentified gold farmers in the data (Ahmad, Keegan, Srivastava, Williams, & Contractor, 2009). Given the presence of identified gold farmers, unidentified gold farmers, and non-gold farmers, we identify three distinct types of character classes. Farmers are the characters whose accounts have been identified as gold farmers by the administrators at any point in time and the trade relationships among them. Affiliates are the farmers as well as any characters that have ever transacted with a character whose account was later banned as a gold farmer. Finally, the nonaffiliates are the characters that have never interacted with identified gold farmers. All three networks are directed, which means that a tie from actor A to actor B is distinct from a tie from actor B to actor A. Because P*/ERG models become computationally intractable to run as the network size grows above approximately 1,000 nodes, we employ a subset of the transaction data for analysis. Furthermore, we specify four specific subclasses of network relationships to capture unique behavioral patterns. The first is market exchange in which the transaction involved one character sending only currency and the other party

25

sending only an item. This network has 1,022 characters and 768 connections. The second is bartering, in which the transaction involved the exchange of items rather than currency. This network has 1,138 characters and 1,323 connections. The third is donations, in which one character gave the other party currency but nothing was reciprocated in that transaction. This network has 1,519 characters and 1,318 connections. The fourth and final is gifts, in which one character gave another an item but nothing was reciprocated in that transaction. This network is the largest with 5,461 characters and 9,239 connections.

--- Figure 1 about here ---

We also use network data on a drug trafficking ring obtained from a Canadian law enforcement taskforce called Project Caviar to contextualize our findings for the gold farming network (Morselli, 2008; Morselli & Petit, 2007). Although orders of magnitude smaller in network size (N=110, E=295), the Caviar data is also directed longitudinal data of criminals and co-offenders which makes it a direct real-world analogue against which we can compare network statistics and dynamics.

Modeling approach Table X describes the relative fit as measured by maximum likelihood estimation as well as the Akaike information criterion. First, a model with a simple set of parameters is estimated to use as a baseline. Subsequent models are built upon this skeleton and their fit are compared against the baseline. Values closer to 0 indicate superior fit. The change in the level of fit against a baseline (such as a model containing only a single, basic parameter like the number of edges in the network) can be assumed to better predict the observed network. Large differences in fit between a baseline model and a model of interest may be evidence that an effect is present in the latter (Contractor, Wasserman, et al. 2006). We use these as a heuristic to evaluate model fit; small changes in fit indicate a relatively poor model while large changes in fit suggest a better model. The model fits are reported in Table 1.

26

--- Table 1 about here ---

The full model incorporating all of the parameters of interest are the best fitting models across the five networks. These parameters are visualized in Figure 2. We review the substantive interpretation of each of these parameters below wherein transactions refer to the particular relationship (market exchange, bartering, donation, and gifting). Out edge: the tendency for characters to initiate transactions. In edge: the tendency for characters to be the recipients of transactions. Out 2-star: the tendency for characters to initiate many transactions. In 2-star: the tendency for characters to be the recipient of many transactions. Mixed 2-star: the tendency for characters to be brokers between other characters. Reciprocity: the tendency for characters to initiate transactions with characters they previously received transactions from. Cyclicality: the tendency for characters to engage in indirect exchange or generalized reciprocity. Transitivity: the tendency for characters to be embedded within hierarchical relationships --- Figure 2 about here ---

Results The eight parameters estimated for each of the five networks had varying levels of valence and significance across the networks. This analysis is primarily motivated by Faust and Skvoretz (2002) definition that networks are similarly structured to the extent that they exhibit the same structural tendencies, to the same degree. As such, we are primarily interested in identifying which pattern of behavior in the four types of gold farming networks generate networks that should exhibit similar structural tendencies as the offline CAVIAR drug trafficking network. By visual inspection alone, we see that the in-game gifting network exhibits the same pattern of valence and similar types of significance as

27

the CAVIAR network. This suggests that the gifting in online transactions exhibit similar structural tendencies as we observe for an offline drug trafficking operation. We will elaborate on these in the context of the gifting network and then highlight the differences in the other relationship classes from this seemingly ideal type below. First of all, links are relatively rare in these networks. The negative estimates for in-edge and out edge suggest that links occur much less often than is to be expected by random chance. This comports well with both theoretical and intuitive notions that people choose their trade partners selectively rather than randomly. However, the positive estimates for the in-two-star and out-two-star parameters suggest that, net of the effect that ties are relatively rare in the network to begin with, there are substantially more instances of characters engaging in multiple trade incoming and outgoing gifting transactions than one would expect by chance alone. Reciprocity also plays a similarly strong effect in both networks; the large estimate suggests that the existence of a tie between characters almost always results in the other character reciprocating. Across the four in-game trade relationship types as well as the CAVIAR trafficking network, the negative mixed-two-star parameter implies that actors engaged in brokering transactions are relatively rare. This is an unsurprising result for the simple fact that being a broker is not only risky but it is also an expensive position to maintain. We also find that transitivity plays similarly strong influences in both the gifting and CAVIAR network. This suggests latent specialization or hierarchy among the actors in each as they preferentially form ties with third parties based on the pattern of ties observed with another alter. However, the negative parameters for cyclicality in both the gifting and CAVIAR networks where strong positive parameter estimates are observed in the other relationships types is doubly notable. First, it suggests that characters in the game trade network as well as drug traffickers avoid engaging in indirect transactions or generalized reciprocity while this structural tendency appears to strongly influence the evolution of the network in the other relationship types. Second, gifting is the type of transaction one might naively expect to see indirect or generalized reciprocity most strongly occurring, but its significant absence for the type of transaction it ostensibly best captures whilst matching the same behavior observed for drug trafficking

28

suggests that the gifting transaction is not really gifting at all. In other words, the movement of items through the network as gifts rather than bartered or market exchanges appears to operate under similar constraints and potentially the same motivations as an offline drug trafficking network. Thus, if one wished to characterize offline drug trafficking behavior by using data from MMOGs, this result suggests that the structure and dynamics of gifting transactions may be the best place to start.

Discussion Recent advances in both statistical techniques for performing multilevel analyses of social networks as well as the growth of virtual worlds as sites for immersive social interaction have opened up new avenues of inquiry for organizational and sociological research. The processes that undergird assembly and maintenance of covert organizations have eluded scholarly attention for decades by virtue of the difficulty of collecting both rich and complete data on actors attributes, interactions, and embeddedness. Moreover, the tiny amount of work done to date has tended to emphasize analyses that failed to account for network processes occurring at other levels of analysis. The framework laid out here offers the potential to develop comparative statistical models of how individual level attributes and structural tendencies influence the interaction patterns on covert networks in both online and offline contexts. Although the evidence for mapping online behaviors, motivations, and affordances to offline contexts is mixed, in some task contexts such as gold farming and drug trafficking, the analogues appear valid and worthy of further analysis. Despite the surfeit of data available from virtual worlds databases which we have extolled repeatedly throughout this paper, P*/ERG models are immensely computationally intensive above networks of approximately 1,000 actors (S. Goodreau, 2007), much less the tens or hundreds of thousands of actors in many games which would demand peta-scale level computing power (Poole, 2009). Approaches to graph sampling are imperfect and ultimately undermine the original motivation to understand large-scale behavior without resorting to estimates, sampling, and surveys.

29

There are substantial ethical implications that accompany theorizing about how to leverage large aggregations of unobtrustive data with indistinct consent and privacy (Lazer, et al., 2009) as well as developing theories and methods to identify and destroy organizations (Canter, 2000). It is certainly the responsibility of researchers to be reflexive about the normative judgments made by political and institutional actors about other organizations. Grounded ethnographic work is likewise needed to more fully explain actors motivations and attitudes as well as the contexts in which they perform these activities that may undermine prevailing narratives about these being deviant entities to be controlled or destabilized.

Acknowledgements This research was supported by the National Science Foundation via award number IIS-0729505, the Army Research Institute via award number W91WAW-08-C-0106, and the Air Force Research Laboratory by contract number FA8650-10-C-7010. The anonymized data used for this research was provided by Sony Online Entertainment. We thank members of the Virtual Worlds Observatory team for their support and feedback.

30

Figures

Market exchange

Bartering

A Donation

A Gifting
Figure 1: Four relationship subclasses

Parameters Model 1 Model 2 Model 3 Out edge + In edge + Out 2 Star + In 2 Star Model 1 + Mixed 2 Star + Reciprocity Model 2 + Transitivity + Cyclicality (Study 1)

Barter -6658.91 13323 -816.47 1645.0 -799.79 1615.6

Market -6278.016 12526 -6188.07 12388 -6178.40 12373

Donate -11129.53 22267 -9278.22 18568 -8798.86 17614

Gift -81928.57 163865 -62132.38 124277 -59382.71 118781

CAVIAR -990.82 1989.7 -714.24 1440.5 -677.01 1370.6

Table 1: Model fitness for five networks. Maximum likelihood estimate (MLE) on top, Akaike information criteria (AIC) on bottom.

31

istar1 istar2 ostar1 ostar2 m2star mutual ctriple ttriple

Est. -7.357 0.913 0.853 -0.324 -2.159 12.681 2.699 -0.094

Barter SE 2.52E-01 1.41E-01 0.00E+00 1.72E-01 2.88E-02 1.09E-01 1.07E+01 9.43E-02

P *** *** *** . *** ***

Est. -7.075 0.126 -0.039 0.181 -0.314 3.524 2.843 1.228

Market SE 4.89E-02 4.58E-02 4.89E-02 4.01E-02 4.76E-02 3.93E-01 1.12E+00 6.84E-01

P *** ** *** *** *** * .

Donations Est. SE -6.786 4.57E-02 0.090 4.57E-02 -0.122 4.57E-02 -0.461 6.91E-02 -0.480 4.46E-02 6.408 6.32E-01 0.147 8.56E-01 3.004 2.71E-01

P *** * ** *** *** *** ***

Est. -8.811 0.135 -0.040 0.074 -0.035 7.195 -0.642 1.359

Gifting SE 1.83E-02 3.77E-04 1.83E-02 1.30E-02 2.12E-03 6.06E-02 2.07E-02 1.25E-02

P *** *** * *** *** *** *** ***

Est. -5.559 0.044 -0.002 0.099 -0.012 4.163 -0.533 0.434

Caviar SE 1.35E-01 2.86E-02 0.00E+00 2.07E-05 4.40E-05 1.31E-03 6.98E-05 4.77E-05

P *** *** *** *** *** *** ***

32

Bibliography

Ahmad, M. A., Keegan, B., Srivastava, J., Williams, D., & Contractor, N. (2009). Mining for Gold Farmers: Automatic Detection of Deviant Players in MMOGs. Paper presented at the Proceedings of IEEE SocialCom 2009, Vancouver, Canada. Anderson, C. J., Wasserman, S., & Crouch, B. (1999). A p* primer: logit models for social networks. Social Networks, 21(1), 37-66. doi: Doi: 10.1016/s0378-8733(98)00012-4 Ayling, J. (2009). Criminal organizations and resilience. International Journal of Law, Crime and Justice, 37(4), 182-196. Bainbridge, W. (2007). The scientific research potential of virtual worlds. Science, 317(5837), 472. Baker, W., & Faulkner, R. (1993). The social organization of conspiracy: Illegal networks in the heavy electrical equipment industry. American Sociological Review, 58(6), 837-860. Balkundi, P., & Harrison, D. (2006). Ties, leaders, and time in teams: Strong inference about network structure's effects on team viability and performance. Academy of Management Journal, 49(1), 49. Bavelas, A. (1950). Communication patterns in task-oriented groups. Acoustical Society of America Journal, 22, 725. Borgatti, S., & Foster, P. (2003). The network paradigm in organizational research: A review and typology. Journal of Management, 29(6), 991. Brass, D., Galaskiewicz, J., Greve, H., & Tsai, W. (2004). Taking stock of networks and organizations: A multilevel perspective. Academy of Management Journal, 47(6), 795817. Brzezinski, M. (2002). Re-engineering the drug business. New York Times Magazine, 23, 24-29. Canter, D. (2000). Destructive organisational psychology. In D. Canter & L. Alison (Eds.), The Social Psychology of Crime: Groups, Teams and Networks, (pp. 321334): Dartmouth. Castronova, E. (2005). Synthetic Worlds: The Business and Culture of Online Games. Chicago: University of Chicago Press. Castronova, E. (2006). A cost-benefit analysis of real-money trade in the products of synthetic economies. Info, 8(6), 51-68. Castronova, E., Williams, D., Shen, C., Ratan, R., Xiong, L., Huang, Y., & Keegan, B. (2009). As real as real? Macroeconomic behavior in a large-scale virtual world. New Media & Society, 11(5), 685. Coles, N. (2001). It's Not What You Know--It's Who You Know That Counts. Analysing Serious Crime Groups as Social Networks. British Journal of Criminology, 41(4), 580. Consalvo, M. (2007). Cheating: Gaining advantage in videogames. Cambridge, MA: The MIT Press. Contractor, N., Wasserman, S., & Faust, K. (2006). Testing Multitheoretical, Multilevel Hypotheses about Organizational Networks. Academy of management review, 31(3), 681703. Cressey, D. (1971). Organized crime and criminal organizations: Heffer. Davis, R. (1981). Social network analysis: An aid in conspiracy investigations. FBI Law Enforcement Bulletin, 50(12), 11-19. Dibbell, J. (2003). Owned!: Intellectual property in the age of dupers, gold farmers, eBayers, and other enemies of the virtual state. In J. M. Balkin & B. S. Noveck (Eds.), The State of Play: Law, Games, and Virtual Worlds. New York, NY: New York University Press. 33

Dibbell, J. (2006). Play Money: Or, How I Quit My Day Job and Made Millions Trading Virtual Loot. New York: Basic Books. Dorn, N., Oette, L., & White, S. (1998). Drugs importation and the bifurcation of risk: capitalization, cut outs and organized crime. British Journal of Criminology, 38(4), 537. Erickson, B. (1981). Secret societies and social structure. Social Forces, 60(1), 188-210. Faust, K., & Skvoretz, J. (2002). Comparing networks across space and time, size and species. Sociological Methodology, 32, 267-299. Frank, O. (1981). A survey of statistical methods for graph analysis. Sociological Methodology, 12, 110-155. Frank, O., & Strauss, D. (1986). Markov graphs. Journal of the American Statistical Association, 81(395), 832-842. Goodreau, S. (2007). Advances in exponential random graph (p*) models applied to a large social network. Social Networks, 29(2), 231-248. Goodreau, S. M. (2007). Advances in exponential random graph (p*) models applied to a large social network. Social Networks, 29(2), 231-248. Granovetter, M., & Swedberg, R. (1992). The sociology of economic life: Westview Press Boulder, CO. Granovetter, M. S. (1973). The strength of weak ties. American journal of sociology, 78(6), 1360. Guetzkow, H., & Simon, H. (1955). The impact of certain communication nets upon organization and performance in task-oriented groups. Management science, 1(3), 233250. Handcock, M. S. (2003). Statistical models for social networks: Inference and degeneracy. In R. Breiger, K. Carley & P. Pattison (Eds.), Dynamic Social Network Modeling and Analysis (pp. 229240). Washington, D.C.: National Academies Press. Handcock, M. S., Hunter, D. R., Butts, C. T., Goodreau, S. M., & Morris, M. (2008). statnet: Software tools for the representation, visualization, analysis and simulation of network data. Journal of Statistical Software, 24(1), 1548. Heeks, R. (2008). Current Analysis and Future Research Agenda on Gold Farming: Real World Production in Developing Countries for the Virtual Economies of Online Games: Institute for Development Policy and Management, University of Mancester. Holland, P., & Leinhardt, S. (1970). A method for detecting structure in sociometric data. American journal of sociology, 76(3), 492-513. Huang, Y., Zhu, M., Wang, J., Pathak, N., Shen, C., Keegan, B., . . . Contractor, N. (2009). The Formation of Task-Oriented Groups: Exploring Combat Activities in Online Games. Paper presented at the Proceedings of IEEE SocialCom 2009, Vancouver, Canada. Huffaker, D., Jing, W., Treem, J., Ahmad, M. A., Fullerton, L., Williams, D., . . . Contractor, N. (2009). The Social Behaviors of Experts in Massive Multiplayer Online Role-Playing Games. Paper presented at the Proceedings of IEEE SocialCom 2009, Vancouver, Canada. Hunter, D. R., Goodreau, S. M., & Handcock, M. S. (2008). Goodness of fit of social network models. Journal of the American Statistical Association, 103(481), 248-258. Jones, C., Hesterly, W., & Borgatti, S. (1997). A general theory of network governance: Exchange conditions and social mechanisms. Academy of management review, 22(4), 911-945.

34

Kilduff, M., & Krackhardt, D. (2008). Interpersonal Networks in Organizations: Cognition, Personality, Dynamics, and Culture: Cambridge University Press. Knoke, D., & Kuklinski, J. (1982). Network analysis: Sage Publications, Inc. Krackhardt, D. (1999). The ties that torture: Simmelian tie analysis in organizations. Research in the Sociology of Organizations, 16(1999), 183-210. Lastowka, F. G., & Hunter, D. (2006). Virtual Crime. In J. M. Balkin & B. S. Noveck (Eds.), The State of Play: Law, Games, and Virtual Worlds. New York: New York University Press. Lastowka, G. (2006). ID theft, RMT & Lineage. TerraNova, (July 5, 2006). Retrieved from http://terranova.blogs.com/terra_nova/2006/07/id_theft_rmt_nc.html Lazer, D., Pentland, A., Adamic, L., Aral, S., Barabasi, A., Brewer, D., . . . Gutmann, M. (2009). Life in the network: the coming age of computational social science. Science (New York, NY), 323(5915), 721-723. Lehtiniemi, T. (2007). How big is the RMT market anyway. Virtual Economy Research Network(March 2). Lindelauf, R., Borm, P., & Hamers, H. (2009). The influence of secrecy on the communication structure of covert networks. Social Networks, 31(2), 126-137. McIllwain, J. (1999). Organized crime: A social network approach. Crime, Law and Social Change, 32(4), 301-323. McIntosh, M. (1975). The organisation of crime: Macmillan. Milward, H., & Raab, J. (2006). Dark networks as organizational problems: Elements of a theory. International Public Management Journal, 9(3), 333-360. Monge, P., & Contractor, N. (2003). Theories of communication networks: Oxford University Press, USA. Moody, J. (2001). Race, school integration, and friendship segregation in America. American journal of sociology, 107(3), 679-716. Morselli, C. (2008). Inside Criminal Networks: Springer. Morselli, C. (2010). Assessing Vulnerable and Strategic Positions in a Criminal Network. Journal of Contemporary Criminal Justice, 26(4), 382. Morselli, C., & Giguere, C. (2006). Legitimate Strengths in Criminal Networks. Crime, Law and Social Change, 45(3), 185-200. Morselli, C., Gigure, C., & Petit, K. (2007). The efficiency/security trade-off in criminal networks. Social Networks, 29(1), 143-153. Morselli, C., & Petit, K. (2007). Law-Enforcement Disruption of a Drug Importation Network. Global Crime, 8(2), 109 - 130. Newman, M. (2003). The structure and function of complex networks. SIAM review, 45(2), 167256. Pattison, P., & Wasserman, S. (1999). Logit models and logistic regressions for social networks: II. Multivariate relations. British Journal of Mathematical and Statistical Psychology, 52(2), 169-193. Pinto, J., Leana, C., & Pil, F. (2008). Corrupt organizations or organizations of corrupt individuals? Two types of organization-level corruption. The Academy of Management Review (AMR), 33(3), 685-709. Poole, M. (2009). Collaboration, integration, and transformation: Directions for research on communication and information technologies. Journal of Computer Mediated Communication, 14(3), 758-763.

35

Raab, J., & Milward, H. (2003). Dark networks as problems. Journal of Public Administration Research and Theory, 13(4), 413. Robins, G. (2009). Understanding individual behaviors within covert networks: the interplay of individual qualities, psychological predispositions, and network effects. Trends in Organized Crime, 12(2), 166-187. Robins, G., Pattison, P., Kalish, Y., & Lusher, D. (2007). An introduction to exponential random graph (p*) models for social networks. Social Networks, 29(2), 173-191. doi: DOI: 10.1016/j.socnet.2006.08.002 Robins, G., Pattison, P., & Wasserman, S. (1999). Logit models and logistic regressions for social networks: III. Valued relations. Psychometrika, 64(3), 371-394. Robins, G., Snijders, T., Wang, P., Handcock, M., & Pattison, P. (2007). Recent developments in exponential random graph (p*) models for social networks. Social Networks, 29(2), 192215. doi: DOI: 10.1016/j.socnet.2006.08.003 Simmel, G. (1906). The sociology of secrecy and of secret societies. American journal of sociology, 11(4), 441-498. Snijders, T. A. B. (2002). Markov Chain Monte Carlo Estimation of Exponential Random Graph Models. Journal of Social Structure, 3(2). Snijders, T. A. B., Pattison, P. E., Robins, G. L., & Handcock, M. S. (2006). New specificiations for exponential random graph models. Sociological Methodology, 36(1), 99-153. Sparrow, M. (1991). The application of network analysis to criminal intelligence: An assessment of the prospects. Social Networks, 13(3), 251-274. Tyren. (2006, December 21, 2006). World of Warcraft Accounts Closed Worldwide, from http://forums.worldofwarcraft.com/thread.html?topicId=59377507 Vaughan, D. (2002). Criminology and the sociology of organizations. Crime, Law and Social Change, 37(2), 117-136. doi: 10.1023/a:1014515700746 von Lampe, K., & Johansen, P. (2003). Criminal Networks and Trust. On the importance of expectations of loyal behaviour in criminal relations. Paper presented at the Organised Crime, Trafficking, Drugs. Waring, E. (2002). Co-offending as a network form of social organization. In E. Waring & D. Weisburd (Eds.), Crime and social organization (pp. 31-48). New Brunswick, New Jersey: Transaction Publishers. Wasserman, S., & Faust, K. (1994). Social network analysis: methods and applications. New York: Cambridge University Press. Wasserman, S., & Faust, K. (1994). Social Network Data Social Network Analysis: Methods and Applications (pp. 28-66). Cambridge: Cambridge University Press. Wasserman, S., & Pattison, P. (1996). Logit models and logistic regressions for social networks: I. An introduction to Markov graphs andp. Psychometrika, 61(3), 401-425. Wasserman, S., & Robins, G. (2005). An introduction to random graphs, dependence graphs, and p*. In P. J. Carrington, J. Scott & S. Wasserman (Eds.), Models and methods in social network analysis (pp. 148161). New York: Cambridge University Press. Williams, D. (2010). The mapping principle and a research framework for virtual worlds. Communication Theory, 20, 451-470.

36

The Game Behind the Video Game Trends in the video games software industry. Toward new business models?

Giuditta de Prato (IPTS), Claudio Feijoo (Universidad Politecnica de Madrid), Jean Paul Simon (IPTS)

Abstract A description of the structure of the global video market together with an assessment of its economic weight is proposed first, followed by a review of its value chain, in order to highlight the evolution of the relationships between the different players of the value chain. Then, the likeliness that disruptive technologies will emerge in the online and mobile games market is investigated, in a scenario of rapidly increasing convergence of digital technologies and integration of media services.

Key words: digital economy, value chain analysis, assessment of the value added, media and content industry new business models, virtual items.

Introduction This paper is based on a comprehensive survey of the video games software industry with a focus on the EU comptetitiveness1, and concentrates on one aspect of this research. The investigated research question is how and to what extent new channels of distribution such as on line games and mobile transformed the structure of the industry with connectivity becoming permanent, products converting into (on-line) services, and the progressive integration of media services and technologies. This paper is organised as follows. Section 1 introduces the available data on the economic dimension of the market. Section 2 investigates the roles of different actors in the value chains and business models on which revenue schemes are based, it stresses the essential role of hardware owners in the industry. It also emphasized the less noted role of middleware. Finally, section 3 examines the impact on on-line and mobile distribution of video games. The research is based on our syntheses of the current state of the knowledge, on internal and external expertise, literature reviews and desk research, as well as several validation workshops with industry participants and experts. 1. The global video games market In less than 40 years, software games developed from scratch into an industry producing billions of profits and today, its revenues and investments give the video games industry a relevant position among other mainstream media industries (Deuze et al., 2007). OECD (2005) mentioned that US games revenue in 2001 had already surpassed that of film box office tickets. In the UK, the video game market outgrew the cinema market in 20092 and playing games online is now as popular as downloading music and video.3 1.1 The size of the global video game market Figure 1 shows estimates of the total size of the global video game market up until 2013. In the period 2004 to 2013, the global video game market is expected to grow from less than 30 billion to over US$ 70 billion.

Figure 1: Global video games market, million US$, PWC 2009

The report Born digital/ Grown digital. Assessing the future competitiveness of the EU video games software industry was released in November 2010. This report reflects the findings of our study on the video games industry, with a focus on two specific activities: on-line and mobile videogames. 2 Cinema market: theatres and DVD. BBC News, Wednesday 24 March 2010: " Rupert Clark, an analyst from consulting firm Deloitte, said that the global games industry now makes more money than the box office" 3 Ofcom Communications Market Report, August 2010. Available at: http://stakeholders.ofcom.org.uk/binaries/research/cmr/753567/CMR_2010_FINAL.pdf

Other sources report similar values of the size of the video games market. For the US market alone, the combined computer games and video games sales in 2007 accounted for US$9.5 billion, an increase of 28% against 2006 (ESA, 2008). According to IDATE (2008), the size of the global market for video games in 2009 was estimated at 46 billion. Developed regions such as Europe, the US and Japan are the main markets for video games. The Europe-Middle East-Africa (EMEA)4 region is the biggest market for video games: France, Germany, Italy, Spain and UK, in 2009, accounted for nearly 30% of the global video games market. It is estimated that in 2009 these regions accounted for over one half, or 26 billion, of the video games market. 1.2 The relative importance of the video game market within the media and entertainment market Table 1 presents the value of the video games market against the background of the entire media and entertainment market.5 According to these figures, in 2007, the global value of video games sold worldwide was US$ 43.5 billion. The growth dynamics forecasted for the video games sector are exceptional, when compared with the entire market for media and entertainment products and services. The former is expected to grow by almost 70% to over US$ 70 billion by 2013, whereas the latter is expected to grow by only 17%. In 2009, the video games market accounted for around 3% of the media and entertainment market. In comparison, the recorded music and filmed entertainment represented 2% and 6% of the global media and entertainment market respectively. In spite of being still a rather young industry, video games already managed to become a significant and growing share of the media and content industries.

4 5

Europe is the core market of this region. Media & Entertainment includes: internet access fees, internet advertising, TV fees, TV advertising, recorded music, filmed entertainment, video games, consumer magazine publishing, newspaper publishing, radio, book publishing, and business-to-business publishing.

Table 1: Global video games and global media and entertainment market, 2007-2013, million US$, PWC 2009
2007 2008 2009 2010 2011 2012 2013

Total Video Games

43,460

51,390

55,089

58,383

61,604

67,026

73,513

Total Media & Entertainment*

1,373,941

1,408,950

1,354,068

1,359,495

1,411,788

1,506,409

1,613,173

* Media & Entertainment includes: internet access fees, internet advertising, TV fees, TV advertising, Recorded music, Filmed entertainment, Video games, Consumer magazine publishing, Newspaper publishing, Radio, Book publishing, Business-to-business publishing

1.3 The value of hardware and software in video games Figure 2 reports the value of software and hardware for consoles and handheld video games and the value of software for the remaining product segments, i.e. PC and wireless. The value of home consoles was around 10,258 million, whereas the value of software dedicated to this platform exceeded 12,559 million. It is estimated that this gap will increase even further and, by 2012 the value of console hardware will drop to around 6,291 million, whereas the value of software is expected to increase to over 16,587 million. Figure 2: Value of hardware and software in video games, by segment, 2008 2012, in million euro, IDATE 2008

2. The value chain in the video games industry 2.1 The main platforms 2.1.1 PC games platforms When taking into account the video games industry as a whole, the PC-based market is smaller than the mainstream one, represented by that of consoles. Nevertheless, it holds some peculiar characteristics, mostly related to the fact that it is on this platform that imaginative programming and risk-taking perform best (Williams, 2002). Figures on the dimension of the PC games market in terms of units of games sold are necessarily more difficult to collect, due to the much bigger number of producers, to the loose linkage with the hardware architecture, and to the much more fragmented market in general. A list of PC game titles which sold at least 1 million copies comprises 96 PC games, without taking into account different sub-releases of the same game. The first three of them, namely The Sims, The Sims 2, and StarCraft, shipped respectively 16 million units, 13 million units, and 11 million units. It can be easily understood that products and economics differ from those of consoles and handhelds segments: in the case of the PC, a common standard is available for the architecture, which can also host different peripherals and powerful additional devices. Microsoft still holds the biggest share of operating systems installations. This implies that third party hardware manufacturers on one side, and independent developers on the other, have a well known and rather stable environment in which to operate. The PC business context reflects low entry barriers which are free from proprietary restriction (Williams 2002) and manufacturers licensing fees, and benefit from lower development costs (no need for specific and highly expensive- software development kits, very low costs of duplication and deployment). These conditions allowed for the exponential increase of game titles: Williams (2002) reports that, in 1998, 4,704 PC titles were available, versus only 44 for the Nintendo 64 (handheld) and 399 for the PlayStation console (NPD Group data). 2.1.2 Console and handheld games platforms Consoles and handheld game platforms are currently the best known set of products in the video games industry, with console products such as the Playstation (Sony), the Xbox (Microsoft), the Wii (Nintendo) and handheld devices such as the Nintendo Ds and PlayStationPortable. This market is dominated by the almost monopolist Nintendo, by means of the long lasting sales success of the GameBoy. By the end of March 2009, 878 million units of GameBoy and its variations had been sold. For the current generation Nintendo DS handheld devices, 596 million were sold. The number of units sold for the recent Nintendo DS are 101.78 million (Nintendo, 2009). Supply has always been highly concentrated in a very small number of producers. As regards consoles, the situation is similar.The total number of home consoles sold from 1977 to date is estimated at more than 664 million units. Even though the number of actors is a higher than that of handheld manufacturers, the supply is still very concentrated. 5

6 In terms of structure of the market and of behaviour of the supply actors, console and handheld systems show pretty similar situations and dominant players. Technological achievements, gaming diffusion across ages and competitive pressure have changed the market. However, high entry barriers still limit the competition in the handheld market to two big players: Nintendo and Sony. In the console market, the main actors are basically three: Microsoft joins the two other giants of handheld devices games. The oligopolistic position of companies in handheld and console segments is evident from the above tables and figures, and is frequently reported in the literature. The reasons can be identified as the high market entry costs related to technology, distribution and the investment needed to develop prototypes. This role of the console is all the more important as the console is the platform managing the network effect in this two-sided market (connecting the two groups of users: developers and players) (Bounie, D., Bourreau, M., 2008). Besides offering a choice of successful games, key aspects to maintain the leadership seem to have been the capacity to reduce costs, successful design, availability of best developers and publishers, high sales capacity, etc. Applications for consoles and handhelds used to share a characteristic which differentiated them from PC applications: it used not to be possible to retroactively fix bugs in software applications for console and handheld devices by means of patches, while this was common in the case of PC applications. This implied that imperfect products needed to be returned. The diffusion of downloadable games has partly solved this limitation, which in the past represented the source of high risks and costs to producers. The most relevant fact in both markets is related to the proprietary characteristics of the devices: each manufacturer defines the technical features and characteristics of its device and the technologies adopted, and, due to the quasi monopoly it holds, is able to impose its solution. The manufacturers control the decision about allowing external developers ("third parties") to develop applications for their devices and hence a common standard is lacking. This, in turn, makes platform interoperability and portability of applications impossible. 2.1.3Mobile platforms The demand for mobile-based video games is relatively new and rapidly evolving, being represented by users of mobile phones and, most of all in recent years, of smartphones. Games started to appear on mobile handsets (mobile phones) about a decade ago (Nokia started installing Snake in 1997), and did not at first raise much interest. In recent years, the rise in the number of developers has been much faster, following the creation of specific mobile game subsidiaries by traditional video game publishers. The investment by telecoms operators has intensified as well, and has differentiated in a number of business activities (publishing, aggregation, distribution, platforms, licensing, etc.). Telecom operators are maintaining the privileged position of being favourite gatekeepers for provision of services to customers, and are preserving their revenues by updating business models towards mobile business. To sum it up, to invest in the sector, high financial capacity is required. Growth of this sector is conditioned by tight technical constraints concerning latency, speed and handset capacity in terms of storage, computation and display.

7 2.2 The value chain in the video games industry 2.2.1 The main actors of the value chain A video game developer is a company that invents and develops video games, and in particular develops the necessary software to run the video game. A video game developer may specialize in a specific video game console, or may develop for a variety of platforms including the PC or the mobile platforms.6 It can also specialise in certain types of games. Developers are usually studios, with multidisciplinary teams. Such companies are small and numerous. In Europe, a large population of these highly creative small development studios is found mainly in the UK, France, Germany, the Nordic countries and to a lesser extent in Spain. Taking into account the specific relation of developers to publishers (see below), and the existence of independent7 developer companies, some developers publish their own games and therefore can be regarded as publishers and developers. This is, for example, the case for the majority of the Norwegian developers.8

A video games publisher is a company that publishes video games that it either develops internally or has ordered from a video games developer. The publisher is responsible for licensing the rights and the concept on which the game is grounded, for handling the marketing and often even the distribution. While the gatekeeper role is played by several hardware platform owners, publishers rarely specialise in only one platform. They opt for platform diversification, but this strategy has its own limits as often titles released for one platform are not compatible with another. Video games distributors usually market the games, handle the packaging and transport, organise the infrastructure for distribution, and sometimes even provide user support. Together with the retailers, they cover the logistics of the chain. Though they are not the publishers themselves, they are usually specialised distributers for video games (and often other digital products). In particular, as large publishers are primarily interested in promoting their own games, independent game companies find small specialised distributors for their titles. There are also large international collaboration agreements such as those for Sony and Nintendo, in which the games and hardware are handled respectively by Nordisk Film (DK) and Bergsala (SW) in Scandinavian countries. Retailers are usually electronic chains, multimedia shops and specialist shops but nowadays video games can be easily found in ordinary distribution stores such as FNAC, Wal-Mart, the Metro group or even Carrefour.

2.2.2 The traditional value chain Figure 3 shows a simplified and traditional view of the value chain for video games. It looks like a classical and linear retail distribution value chain. The product, from its creation to its
6

In this case, the availability of platform-independent middleware is a key factor in reducing development costs and allowing multi-platform development. Please refer to Chapter 5 for details on layers of software and middleware in particular. 7 Independent companies aim to maintain and grow their business without having to develop games on demand from publishers. 8 See in: Norwegian Ministry of Culture and Church Affairs, 2008. Video games. Report 14 (2007-2008) to the Storting (Norwegian parliament.).

8 consumption goes through a series of necessary intermediaries to allow for its commercialisation, each of the intermediaries exercising its specific role and aiming to optimise its profit and position. This presentation excludes the hardware production part, reserved domain of the console manufacturers dealt with in the previous section. Figure 3 Video games traditional value chain

Obviously, different actors with different objectives and competences are occupying the various positions in the value chain. Their mutual relations create the value chain dynamics, and shed light on the potential transformations that this value chain might incur in the case of disruptive trends. The distribution of revenues between the stakeholders is difficult to apprehend. According to Wi, J.H (2009), in the case of an off-line game where the retail price is US$58, the retailer will keep US$18, the wholesaler US$10, and production cost will require another US$10, which leaves a profit of US$20 for the developer. 9 In the traditional value chain, the platform hardware owners (Sony, Nintendo, Microsoft) develop, within a strongly oligopolistic market, strategies reinforced by the proprietary characteristics of the operating systems and by the vertical integration of the industry. This dominant position creates tensions with the complementary need to develop an active community of developers, usually small studios, gathering multidisciplinary teams around the creation of the games. Publishers used to occupy a position of strength, partly because the production of video games and all digitalised creative content goods is characterised by high fixed costs and low marginal costs. With further digitalisation and standardisation, the balance of power between these actors could be affected. A growing number of new actors is therefore foreseen, which will position themselves in the video games value chain as video games go progressively online and mobile. 2.3 A core technology: the middleware Video games are built upon a (software) game engine, which represents the core technology. The core game engine allows the higher level application (the part of the game containing the content) to more easily interact with the lower layers, drivers and operating system, and, as a consequence, with the hardware. The game engine is meant to be in charge of heavy and
9

At 8. For example, the cost structure a book is split as follows: author (11%), publisher (14%), printer (16%), distributor (wholesale part, 11%: logistics, + "diffuseur" e.g. sale force- : 7%), retailer (36%), VAT (5.5%). Genvo S. and Solinski B. (2010) are offering the following break-up of a 55 euros retail price: game designer: 14%, editor 29%, console manufacturer 22%, retail outlet 35%.

9 repeatedly accessed routines, e.g. it deals with graphics rendering and with the "intelligence" of the game. The engine is in charge of detecting the interaction of entities in the game, the reactions to each action, and so on. Moreover, the middleware provides the developers with an effective development environment. Between the OS and device driver layer and the end-user application, an intermediary layer can be identified, though it is usually hidden from the applications end users. This layer mainly refers to applications like engines, which are meant to be used (called) by higher level applications and are designed to provide them with specific functionality. They usually come in the form of Software Development Kits, which are sets of development tools, including libraries and applications which allow software developers to develop applications faster and in a portable way. The purpose of Software Development Kits and engines in general can be very diverse, and with regard to video games there are several types of engines. The presence of an additional layer, besides that of drivers, is needed for two reasons. Firstly, it ensures the reduction in application development costs by allowing reuse of components on one side, and secondly it improves efficiency and effectiveness in the applications' development process, thus making it possible to develop cross-platform applications. The typical duration of the whole development cycle of applications in general, and games in particular, is usually longer than the applications life. The reduction of developing and testing time is a requirement of highest importance. The possibility for an application to run under different environments, or on different platforms in the case of video games, is generally referred to as portability. In terms of the business model behind game engines distribution, game engines are distributed under different types of licences. Basically engines divide into: i) free open source engines, generally distributed under a licence of the GPL (General Public Licence ) type together with the source code (along with the open source approach), ii) freeware engines, distributed freely but without the source code, and iii) commercial engines, which are proprietary engines distributed under the payment of a royalties or similar commercial contracts. It must be underlined that the third party middleware producer environment is still changing fast. Successful companies, usually founded by a few brilliant young developers, which licence the core middleware portion of their software, are often bought by big studios or platform producers. Then they are either integrated into the buyers company structure to work in-house (as a first-party development entity) or kept as satellites dedicated to second party development. In general, survival as third-party independent development companies is directly connected to the success of the games their modules are built in. It is quite difficult to maintain an updated list of all the available middleware products for video games, nevertheless Figure 4 shows the distribution of companies among European Union countries, other countries and the United States for 212 middleware applications listed by GameMiddleWare.Org.

10

Figure 4: Geographical distribution of middleware producers

Source: IPTS elaboration on information provided by GameMiddleWare.org (accessible online at: http://www.gamemiddleware.org/ (last accessed: 17 December 2009).

The United States provides 93 of the 212 analysed items, 43.9% of the total. The figure shows that 39.6% of modules are produced by companies based in one of the European Union countries10. Of these 84 products, 22 were developed by companies in the UK, 19 by companies in France, 12 in Germany, 5 in the Netherlands, 4 in Ireland and 4 in Sweden. 35 items, or 16.5% of the total, come from other countries. 14 of these have been developed by firms in Canada. 3. The coming of an era: online and mobile games Several trends are expected to affect the current and future dynamics of the video games software industry. Since 2004, the online and wireless market has grown with remarkable rapidity, driven by the increase in the number of broadband subscribers, the innovation in available games, and the transition to handheld devices and the newest generation consoles (PWC 2009). The sales of online and wireless video games have been increasing rapidly over the last few years and are likely to grow further in the foreseeable future. This shows that the dynamics of this market are driven by technological novelties and new applications. Mobile games are challenging the monopolies of existing operating system owners and are offering a new distribution channel to developers. Moreover, online games (Massively Multiplayer Online Games - MMOGs - and also easier to play browser-based, single user online games) are offering users a new role, which could bypass the publishers and create different revenues streams. 3.1 Online value creation Online games benefit most from the increase in electronic diffusion of content, as this guarantees, on the one hand, growth in the number of users and demand, and on the other
10

Middleware from Unity3d (a Danish firm) is used by 10 to 20% of the top 100 games. Steffen Toksvig, Unity3d, presentation at the IPTS workshop, June 2010.

10

11 hand, allows games to be improved by adding new functionalities, and opening new subsegments as it becomes possible for millions of players to interact simultaneously. Video games, which normally allow a non-linear interaction with the user, are making the most out of their online possibilities. Going online enables them to exploit the promises offered by massive multi-player interaction: the creation of persistent virtual worlds and characters, multiple entry points and continuously updated plots enriched by the inclusion of user-determined content. Independent application stores and console-oriented gateways are differentiating the type of digital content, resources and services made accessible to users. The progressive but impressively fast switch to online gaming introduced new distribution methods and started to rearrange the relative roles and interaction dynamics among the actors at the different levels in the supply chain. Clearly, logistics has lost relevance in the online games segment due to the fact that digital goods are reproduced and distributed over the network at low cost. A whole part of the core business involving publishers, distributors and retailers has basically disappeared as there is no longer any need to duplicate physical products because these can be distributed over the network. The publisher, in many cases, directly distributes games, without the need for a distributor to act as intermediary between the publisher and the retailer: i.e. "disintermediation" is taking place, cutting out the role of the distributor.11 Publishers can also opt to distribute games through Internet Service Providers (ISPs). ISPs act as content aggregators and provide portals for game distribution which allow easier promotion and localisation of new games by users; at the same time they attract advertising which brings an added source to the mixed revenue models. The increasing importance of ISPs has triggered a process that is often labelled "re-intermediation": ISPs are taking on the role previously played by distributors. In the past years, the distribution of online games has been progressively concentrated on some internet portals serving the PC-based side (like, among many others, Valve's Steam Service or Manifesto Games), and on a few, very powerful, network platforms for console games, each controlled by the console's hardware provider. Independent application stores are growing rapidly,12 providing online games access to PC users together with the possibility to download not only games but also movies, music, and additional content. In the same way, console-oriented gateways are also increasing their importance and audience by differentiating the type of content and services made accessible to users. Having started as gateways for accessing video games, and related contents and communities, they are more and more offering different kinds of digital contents and resources.13 The alternative business models which users face when entering the world of online games are actually rather different from those they were used to. As regards demand, consumers are
11

Disintermediation is also taking place in the case of off-the-shelf games, where the increase in structure and negotiation power of big retail chains has allowed them to interact directly with publishers, leaving distributors with a marginal role. 12 For instance in Germany in 2009, two browser-based game companies (Bigpoint, Gameforge) were among the five fastest growing IT companies of the country. Source: http://www.deloitte.com/view/de_DE/de/branchen/article/5bcc6816ec574210VgnVCM100000ba42f00aRCRD.h tm 13 The key dynamics of video games in general are described in a more general framework in Mateos-Garcia et al. (2008).

11

12 attracted by the free-to-play (F2P) approach to the video game main product, because they see it as less of a financial risk. Users are more confident and more willing to pay small sums for digital items offered to enhance their gaming experience, once they already know the game itself and enjoy playing it. Currently, the emerging revenue stream from selling virtual goods online is attracting a lot of attention in the online video games industry. The virtual items model allows gamers to buy individual digital components such as virtual currency, items, characters, and any in-game good which are not a full game in themselves. 3.2 Going mobile The mobile gaming ecosystem allows three basic mechanisms to deliver and consume mobile games: over mobile telecommunications networks, over some short-range wireless system (context delivery) and over some fixed Internet access and later side-loading. A new approach to mobile games, which completes those of mobile operators, content providers (game publishers) and device suppliers, has appeared with considerable strength: the application stores and the platforms that support each of them. In fact, the impact of new platforms and application stores has been considerable from the perspective of mobile gaming development. While development and marketing costs for a console or PC game can ran to millions of Euros, typical costs for a mobile game were already in the range of the hundreds of thousands, sometimes even less (Soh & Tan, 2008) before the emergence of platforms. In the new platforms, these costs may even be an order of magnitude less.14 Thus, the low entry barriers for mobile games have helped spawn a proliferation of small mobile-game software developers and the possibility to account for the long-tail of potentially interested gamers. The success of mobile gaming is linked with the appearance of consolidated and scalable business models. There are no big surprises in the main business models for mobile games. In the case of games publishers (content providers), they are fundamentally a translation of the existing business models of the software game industry into the mobile domain: retailing (pay-as-you-go), premium retailing (the game with basic functionalities is free) and subscription (for gaming online). The business models for the other main types of players operators, suppliers and application stores owners- derive from the discussion in the proceeding sections and basically rely on their market power in the mobile ecosystem to arrive at some form of sharing revenues with the games publisher, or to benefit from their position (gatekeeper role, as mentioned before) in the ecosystem. With the success of application stores, there has been a recent interest in finding new business models which could build on their particular features. Value-added applications, i.e. applications downloaded from an application store from which -during its use- it is possible to access to new functionalities, provide a way for mobile game developers and publishers to utilise business models which have evolved beyond the traditional pay-per-download to incorporate billing from within the app itself for a variety of additional content and services (Holden, 2009). These can include time-based billing for game subscriptions, event-based billing (subscribing to an event music, video, etc.- through the game) or item-based billing (e.g. payment for an additional level or piece of weaponry on a game). It is expected that this

14

According to Nokia sources, this costs would be in the range of the 10 000 euros in 2010 for an average application

12

13 new type of business model will bring in a relevant part of the overall revenues of the mobile content and applications market The necessary conditions for the success of mobile content and applications seems already met in most developed countries, with broadband mobile data networks increasingly available and affordable, in addition to many features offered by the mobile platform to support a massive adoption of gaming (wide demographics; ubiquity; personal devices linked to social networks, community involvement, co-creators of content, context awareness). However, mobile gaming also faces a number of challenges, ranging from technology and economics to the institutional/ regulatory framework. Conclusion Online and mobile games have a role in the digital content convergence process. This is in line with the process of digital convergence and which is based on digital distribution of different types of content on the one hand, and on the diffusion of the availability of interactive capabilities to the consumers on the other. This phenomenon is not only affecting the video game industry, but also the movie, video, music and mobile communication industries and the whole publishing sector in general. One of the disruptive trends in the video games business is the emergence of new actors from different businesses, which may be able to bypass existing actors in currently dominant positions. New companies, such as online portals (MSN, Google, Yahoo, pogo.com), Internet service providers, online social networks (Facebook) or even telecom operators (Orange, Vodafone) or telecom equipment manufacturers companies (e.g. Nokia) may become essential intermediaries in the video games value chain. These entries will bring new form of intermediation that may or may not be welcomed by incumbent players. At the same time, the technological move toward network gaming is also allowing some disintermediation, as the section on online games has clearly shown. Fewer parties involved in the value chain may mean more revenue for the remaining parties. For instance, developers may benefit from direct contact with the consumers which will make them less dependent on the established publishers. Consequently new business models are emerging, however, it is not yet known which the most successful business models will be and when they will be in place. The case of mobile gaming is paradigmatic. From a historical perspective, there was first a business culture clash between mobile operators and content / applications providers, which has been only lately and partially solved as application stores have burst onto the mobile scene. Problems related to pricing, usability and the processing power of mobile phones were next to be solved with the market success of smartphones and other smart devices. The lack of mobile broadband, which has impeded on-line and social gaming, is also becoming an issue of the past, at least in most parts of the developed world. Notwithstanding the above accomplishments, the echoes of these difficulties surface in the complexities of the mobile ecosystem, in the fight for the control role in the emerging platforms within the ecosystem, and in the different perceptions of mobile game evolution: still for many game developers, mobile devices are just another distribution channel, and as far as the mobile industry is concerned, games are just another type of content/application. The videogames software industry appears to be one of the most innovative labs for the coming Digital Economy: it is developing and experimenting new digital services (on-line, off-line and mobile) that manage to reach a growing share of the population. Born digital, the industry shows a digital growth that is taking advantage of many opportunities to offer user13

14 friendly, intuitive services at a very large scale. Such services, mainly based on software development, are progressively invading other areas in the sector such as casual games,15 advergames16 or edutainement,17 multiplying the supply-side actors.

15 16

Casual game: ease of use games (to learn, to access and to play) spanning all genres. Advergames: a subset of the so-called serious games (i.e. allowing for other uses than entertainment), sponsored and distributed for free to advertise a product or an organisation. 17 Edutainement: games with educational outcomes targeted at specific groups of learners.

14

15 The video games report is part of the European Commission COMPLETE Project (20072010). Objectives: Analyse the future competitiveness of the EU ICT sector in emerging ICT technologies. COMPLETE aims at producing 6 reports, focused on the future industrial European competitiveness in the following emerging technologies: WEB Displays, RFID, Robotics, Video Games Software Embedded Software in automotive Semiconductor design with DG INFSO. 2.0,

COMPLETE is co-financed by JRC-IPTS and DG ENTREPRISE and INDUSTRY of the EC. References Bounie, D., Bourreau, M; (2008), "Les marchs deux versants dans les mdias", Culture Web, Xavier Greffe and Nathalie Sonnac (eds), Dalloz, Paris, Chapter 26, pp. 477-491. De Prato, G. Feijo, C., Nepelski, D., Bogdanowicz, M., Simon, J.P., (2010) Born digital/ Grown digital. Assessing the future competitiveness of the EU video games software industry available at: http://ipts.jrc.ec.europa.eu/publications/index.cfm Deuze M., Bowen M. C., Allen Christian (2007), The Professional Identity of Gameworkers, Convergence: The International Journal of Research into New Media Technologies, Vol 13(4): 335353, Sage Publications, London, Los Angeles, New Delhi and Singapore. ESA Entertainment Software Association (2008), Essential Facts about the computer and video game industry. 2008 Sales, Demographic and Usage Data, ESA. ESA Entertainment Software Association (2009), Essential Facts about the computer and video game industry. 2009 Sales, Demographic and Usage Data, ESA. Feijo, C., Maghiros, I., Abadie, F., & Gomez-Barroso, J. (2009). Exploring a heterogeneous and fragmented digital ecosystem: mobile content. Telematics & Informatics, 26(3), 282-292. Feijo, C., Maghiros, I., Bacigalupo, M., Abadie, F., Compa, R., & Pascu, C. (2010, forthcoming). Content and applications in the mobile platform: on the verge of an explosion. Seville: Institute for Prospective Technological Studies Game Developer Research (2009a), Game Developer Top 20 Publishers 2009 Report.

15

16 Game Developer Research (2009b), Game Developer 2009 Government Game Incentive Report. Game Developer Research (2009c), Top 50 Developers 2009. Genvo S., & Solinski B. (2010), "The video game: a cultural asset ?". INA Global, http://www.inaglobal.fr/en/video-games/article/video-games-cultural-asset Holden, W. (2009). Mobile apps - taking a bit of the apple: Juniper Research. Retrieved from http://www.juniperresearch.com/shop/products/whitepaper/pdf/Juniper%20Research%20%20Apps%20White%20Paper%202009.pdf IDATE (2008), Digiworld Yearbook 2008, IDATE, Montpellier, F. IDATE (2009), Digiworld Yearbook 2009, IDATE, Montpellier, F. In-Stat (2010b), The Digital Entertainment Revolution, White Paper, IN1004828WHT, February 2010, available at www.in-stat.com (last accessed: 12 March 2010). Leadbeater Charles (2008), We-think: Mass innovation, not mass production: The power of Mass Creativity, Profile Book. Mateos-Garcia J., Geuna A., Steinmueller W.E. (2008), The Future Evolution of the Creative Content Industries- Three Discussion Papers, pp.16-17. Fabienne Abadie, Ioannis Maghiros, and Corina Pascu, (Eds). IPTS, Sevilla, p.34, 35. Spain. Norwegian Ministry of Culture and Church affairs, 2008. Video games. Report 14 (20072008) to the Storting. OECD (2005), Digital Broadband Content: The online computer and video game industry. Working Party on the Information Economy, OECD, 12 May 2005, DSTI/ICCP/IE(2004)13/FINAL. OECD. The Future Digital Economy: Digital Content Creation, Distribution and Access. Technical report, 2006. http://www.oecd.org/dataoecd/54/35/36854745.pdf. Ofcom Communications Market Report, August 2010. Available http://stakeholders.ofcom.org.uk/binaries/research/cmr/753567/CMR_2010_FINAL.pdf at:

PriceWaterhouseCoopers (2009), Global Entertainment and Media Outlook 2009-2013, 10th Annual Edition, priceWaterhouseCoopers. Screen Digest Ltd, CMS Hasche Sigle, Goldmedia Gmbh, Rightscom Ltd (2006), Interactive content and convergence: Implications for the information society - A Study for the European Commission DG Information Society and Media, Final Report, October 2006, European Commission DG Information Society and Media. Wi, Jong H. (2009), Innovation and Strategy of Online Games, Imperial College Press. Williams Dmitri (2002). A Structural Analysis of Market Competition in the U.S. Home Video Game Industry. International Journal on Media Management, 4(1), p. 41-54. Williams Dmitri (2003), The Video Game Lighting Rod. Constructions of a new media technology, 19702000, Information, Communication & Society, 6:4, 2003, pp. 523550, University of Michigan, USA.

16

Hacktivism in Online Games: Negotiating Transparency, Privacy, & and Policy Making Peter Ludlow Northwestern University Department of Philosophy Crowe 3-159 1880 Campus Drive Evanston, IL 60208 Burcu S. Bakioglu Indiana University Department of Communication & Culture

800E.ThirdSt. Bloomington,IN47405 Abstract


In our paper, we examine the leaking operations that take place in Massively Multiplayer Online platforms and argue that any case that can be made for protecting whistleblowers in the physical world extends into virtual worlds and online games. Therefore, we interrogate to what extent transparency and privacy could be established under the policies set forth by the current terms of service, and more important, to whom does the responsibility belong in securing personal information and intellectual property.

Introduction
In 2010 Wikileaks blew up in the international media and the subsequent 4Chan Operation Payback helped hacktivism become a household word around the world. Less well known is that hacktivist movements and leaking operations have also taken root in Massively Multiplayer Online Games and other virtual worlds. In some cases, these operations have been part of guild-on-guild psyops, but in a few recent cases the in-game hacktivists have exposed secrets about online organizations, ingame vigilante groups, the game companies and their affiliated service providers. For example, in one recent case, reported in Ludlow (2010) and Jenkins & Ludlow (2010), a 4Chan-related group in Second Life known as The Wrong Hands infiltrated an online vigilante group (JLU) and exposed a wide surveillance operation that this group had been conducting. More important, the leaked evidence suggested that these illegal activities were being backed by Linden Lab (Second Life's parent company), which was providing the group with god-mode capabilities that facilitated the operation. In a later development, The Wrong Hands exposed the shady past of the key programmers of a software company (Modular Systems), which was developing a viewer for Second Life. These programmers, operating under the disguise of white-hat hackers, had backgrounds not only in writing malicious code, but also other activities that violated the terms of service -- in particular, content theft. The Wrong Hands, this time, uncovered an even bigger surveillance/datamining operation conducted through the Modular Systems's viewer. The company, which had several former Linden Lab employees on staff, folded shortly after the expos. Hacktivist movements inside virtual worlds raise a series of interesting questions that have direct relevance to the real world. For instance, they don't just expose shady dealings, but rather, question our understanding of online privacy. Game companies have a responsibility to protect the privacy of their users, but just how far does this responsibility extend? In this paper, we review the above incidents, and then argue that any case that can be made for protecting whistleblowers in the physical world extends into virtual worlds and online games. We also

probe to what extent transparency and privacy could be established under the policies set forth by the current terms of service, and more important, to whom does the responsibility belong in securing personal information.

1.0HacktivismEntersVirtualWorlds
Our story begins with a Second Life user-created vigilante group called The Justice League United (JLU). The JLU was founded by a user named Kalel Venkman on April 4th, 2006. From the beginning, the JLU presented itself as an anti-griefer organization dedicated to keeping peace and order in the virtual world. At the time of the JLUs founding, a griefer group called the Patriatic Nigras (PN), which had originated from the 4Chan message boards, had made a name for itself in SL through its transgressive and scatological behavior. Their activities included grid attacks, feigned racism and intolerance, and persistent harassment of other residents. Kalel Venkman, created his group to fight the likes of the PN. In line with the role-playing that SL has become famous for, Kalel took on the likeness of Superman and named his group the "Justice League United" (JLU) whose members assumed the guise of comic book superheroes such as Green Lantern, Batman, Wonder Woman, etc. To make the fantasy complete, they designed their headquarters with a NASA quality control room with monitors that displayed constant updates from all over SL. The updates also informed the League members what representatives of the game company (Linden Lab) were online. As it later turned out, JLU had close contacts with Linden Lab employees and utilized those relationships in filing abuse reports against other players. Predictably, this seemingly innocuous and heroic role-playing initiative turned sour as the overzealous JLU members began heavily abuse reporting even the unlikely groups in Second Life. As reported by Ludlow (2010), JLU was banned from Furnation (a region in SL where players appear in anthropomorphized animal likenesses) because of their excessive vigilantism. Corsi Mousegard, the owner of the Furnation region in SL was later to explain to Alphaville Herald (2010, September 3) that the JLU had become a bunch of power hungry crazed nutjobs that think being an SL Police gives them the right to get away with whatever they want. Likely, there were economic motivations that drove the aggressive abuse reports filed by the JLU. According to a Intlibber Brautigan, an SL real estate mogul and the co-founder of Proactive Security (a virtual world security business in SL that competed with the JLU for business), JLU had already penetrated the PN, but instead of using these moles to destroy the griefing group, they were organizing them to raid, or attack, newly found estates. These raids justified the existence of JLU as a necessary force in SL and the estate owners were, then, asked to join Proactive Security.1 JLU, in turn, were claiming that Intlibber had hired the PN to grief his enemies in the virtual real estate business. (Ludlow, 2010). To be clear, JLU wasnt just falsely accusing innocent users; they did spar with griefers and fight disruptive behavior. But the issue was to figure out who was the griefer and who was the bystander. For example The Alphaville Herald, which had been reporting on virtual world griefing since 2003, was viewed to be just as culpable of griefing as those who were its subject matter. The editor of the online newspaper, Pixeleen Mistral, was later to find out that she was on the JLU target list. There were other actors who were complicating this play, including a small university from Southern California: Woodbury University (WU). Established in SL in 2006, WU had always adopted an alternative and experimental approach to education. According to Dr. Edward Clift, the Director of Media, Culture, & Design at WU at the time, their campus was not just a static virtual mock-up of their actual campus, but rather, an active, living, breathing entity where its students could study and creatively experiment with these social relationships and online cultures in an educational setting.2 Not surprisingly some WU students embraced the 4Chan

culture and their memes. This in turn led the students to befriend the PNs. Ultimately, Tizzers Foxchase, the chief officer of WU, did the unthinkable. She opened the doors of their land to the PN. Interacting with the PN, who continued their disruptive activities while they were hanging out in the Woodbury region, led the WU to be seen as part of this notorious group of virtual world punks. After the resulting ruckus led to the university losing its educational discount and its students getting mass banned (mostly thanks to the aggressive abuse reports of JLU), the university bought four new regions that the students designed with Soviet themes to tweak a famous Second Life troll with ties to the former Soviet Union. Because these regions had open access (unlike most other educational spaces), a group of hackers, who were developing a viewer (browser) for SL called Emerald, also made the Woodbury region their regular hangout and even had an office there. In the meantime, people were beginning to sense curious connections between Linden Lab and some of the vigilante groups in SL. For example, the JLU appeared to have a hotline to the offices of the Governance Team (G-Team) of Linden Lab and were able to have SL users banned on a whim. Another curious alliance was with a programming company called the Modular Systems, which was developing the Emerald viewer for SL. Even though some of the Emerald developers had a shady past, they quickly became Linden Lab's development team du jour. To be sure, similar alliances between Linden Lab and third party developers had existed in the past, (for example, in the case of libsecondlife developers), and they were viewed with equal suspicion. However, none of the libsecondlife hackers had intentionally been mixed up in criminal activity nor did they engage in black-hat hacking. The same could not be said of some of the members of the Emerald team who had been banned from SL on several occasions for content theft and other crimes against the Linden Lab Terms of Service (ToS) agreement. In 2008, a member of the JLU quit the group and gave an interview to the Alphaville Herald, detailing the operations of the JLU, claiming that they were keeping massive intel on SL users, were abuse reporting people capriciously, sometimes successfully getting them banned without cause, and that employees of Linden Lab were complicit in these operations. These charges were dismissed by the JLU. Nevertheless, fed up with JLUs aggressive abuse reports, WU began to plot an infiltration operation. It was clear to them that something foul was going on, and while they were no angels, the WU group was paying the price. Getting tired of constantly being banned for no reason other than their association with the PN, a group of Woodbury students formed a subgroup called The Wrong Hands. According to Tizzers Foxchase (a.k.a. Jordan Bellino), the chief officer of the Woodbury group in SL, The stated goal of the Wrong Hands was to enlighten the SL community about some of the more scandalous behind the scenes happenings on the grid. And that they did. The Wrong Hands started its operations with JLU in their sights. Early January of 2010, a certain Haruni Thespian volunteered to become a JLU member, and much to WUs surprise, he was granted full access to the BrainiacWiki only after a four-day wait period without a non-disclosure agreement. During that time, Haruhi enacted the role of a young girl who sought the fatherly assistance of Kalel Venkman in her personal affairs (boy troubles in particular) while at the same time reporting her findings to the Skype chat group which was put together for what was soon to be called The Wrong Hands. After infiltrating the group and gaining access to its secret Wiki, Haruhi leaked its 1700-pages of confidential information. Three days after the security breach, Haruhi politely sent a letter to all of the JLU members ironically thanking them for their hospitality. The note even started with a pertinent quote from Supermans arch-nemesis, Lex Luthers: Gods are selfish beings who fly around in little red capes and don't share their power with mankind. No, I don't want to be a *god*. I just want to bring fire to the people. And... I want my cut - Lex Luther Well boys and girls, its been real. I wanted to give you all a nice big thank you on behalf of Woodbury University for the Wiki and all the information your loose lips have slipped. There are a few of you who I would like to thank by name, and they are And thus, the League was officially pwned.

As Second Life denizens and Alphaville Herald reporters dug into the contents of the Wiki, they discovered that it was not only used to track suspected griefers and compile information on SL users, but it also contained records of chatlogs from JLU meetings with unsuspecting parties and it recorded JLU plots to entice the Linden Lab Governance Team into banning players. Once leaked, these papers were posted on at least five file-sharing sites free to download for anyone who was interested. Exposing the inner workings of the controversial anti-griefing group, this leak damaged JLUs reputation and credibility because it became obvious that the papers contained a considerable amount of misinformation and suggested curious links between them and the Linden Lab Governance Team. Upon finding out about the breach, Venkman, frantically threatened to take legal action against not only Haruhi, The Alphaville Herald (SL newspaper), and various bloggers who published the contents of the Wiki that he believed to be copyrighted. Fortunately, Haruhi had kept all the chatlogs to prove that he had done nothing wrong, that he was granted legitimate access to the Wiki, and that he was never asked to sign a non-disclosure agreement. In the face of a DMCA take down, the editor-in-chief of Alphaville Herald, Pixeleen Mistral, filed documents disputing Venkman's claims of copyright infringement on the grounds that the materials were a clear case of fair use for journalism. But in the process, Pixeleen was forced to reveal his real life identity as Mark McCahill, a secret that he had been guarding for six years. When Venkman failed to counter-file, Typepad restored the content it removed from Alphaville Herald's site. In effect, the DMCA take down request was an attempt to cover up the extensive surveillance operation that the JLU was conducting and to keep the SL users from understanding the full scope of its abuses in SL. Whatever JLU's ultimate motives might have been, the incident raised serious questions about the ToS of SL and copyright policies in general. The JLU incident was soon followed by another scandal this one involving some of the developers of a Second Life viewer called Emerald. Linden Lab has had an ongoing policy of working with third party developers who actively reverse engineer the code of Second Life to provide a more flexible, user-friendly environment for its end users. Consequently, there have been several groups of programmers who actively work on developing third party viewers to log onto SL and for the most part they are more functional than the default client that Linden Lab provides. Among these, the Emerald viewer, which at one time boasted of 76,000 unique users, became one of the most popular viewers. Yet the ongoing rumors of questionable conduct on the part of some of the development team, in particular Fractured Crystal (Jcool410), not only led the viewer to be labeled as a griefer client (because it was permissive enough to commit ToS violations), but also led some of the SL users to avoid the client all together. Suspicious of some underhanded dealing that may be going on through this client, The Wrong Hands interrogated Fractured Crystal about the Emerald viewer. The video of the interview, in which Crystal openly admitted to ToS infringements, was posted on YouTube on April 11, 2010. Under the questioning of the Wrong Hands agent Cam Scientist, Crystal confessed to his other alt accounts having been banned as a result of ToS violations (i.e. griefing) and openly agreed that the Emerald viewer had functionalities similar to those of other viewers that were favored by griefers. The only difference between Emerald and other griefer viewers, Crystal noted, was that he and his friends were in charge of this one. Of course, Crystal's shady past made the statement all the more alarming and undermined the credibility of the viewer. This interview, along with Linden Lab's apparent cover up of Crystals past and its support of the viewer, shed a questionable light on the Labs motives. As residents were voicing their reactions to this unexpected cover-up, possibly because of the expos of Crystal on YouTube, the Emerald team closed their bureau in the Woodbury region and pulled out their advertising for their viewer from that region. While some assumed that this was an attempt on the part of the Emerald team to clean up their act, others speculated that Fractured Crystal knew, or even was involved in, the decision to have Woodbury be booted out of SL as revenge for their probing questions. Some even suspected that Linden Lab, for various ideological and pragmatic reasons, seized the moment to use this as an excuse to put one griefer group out of business, and chose to deal with the other at a later date. In retrospect these speculations were not unfounded. April 19, two days prior to the

Woodbury ban, Modular Systems, the company that developed Emerald, rolled out their new viewer compatible with Linden Lab regulations and was approved as one of the official viewers. It wasnt long after Crystal's questioning that the Wrong Hands hacked into the Modular System's database and leaked incriminating documents. Apparently, the Emerald developers had compiled a database of avatar names, IP addresses, and geo-location information for players who created SL accounts at the ModularSystems.com site. The leaked documents included e-mail exchanges, a database containing over 39,000 entries tracking 16,740 SL users, detailing avatar name, avatar key, and IP address, source code for portions of a datamining application, and a picture of the Emerald developers in a meeting with the Linden Lab CEO at the time, M Linden, Linden legal counsel Marty Linden, and several other SL staff. This revelation not only exposed the close relationship between Linden Lab and the Emerald developers (whose reputations had already been tainted), but also raised some serious privacy concerns, concerns which were apparently shared by the Linden Lab developers. Soon enough, chatlogs of Crystal saying that the consequences for the Woodbury faction, as a result of the leak, could be severe, hinting at a possible mass ban. And so during a meeting in which WU was negotiating its legitimate return to SL, their lifeline was unplugged. Woodbury University was permanently banned yet again as they were days away from getting their educational island.

2.0Analysis
These events bring into sharp focus the fragile status of user privacy in online worlds as well as the apparent role that hacktivist action can play in preserving privacy. The paradox, of course, is that sometimes intrusions into private networks by hacking operations can expose and dismantle far more dangerous and systemic violations of privacy by institutions and coding authorities. In the case of the JLU, an intrusion by the Wrong Hands into the private archives of the JLU not only exposed a systemic surveillance operation that was violating the privacy of thousands of unaware users, but it also exposed the involvement of the coding authority (Linden Lab) in the matter. This naturally raises the question of when such operations are justified. Under what conditions are hacktivist operation like those undertaken by The Wrong Hands warranted? Answers like never and always dont seem to be fruitful here, because the actions of The Wrong Hands exposed privacy violations far worse than those they themselves were undertaking. But when exactly are such actions justified? Unfortunately there are no easy answers here. In both the case of their JLU infiltration and their hacking of Modular Systems, The Wrong Hands had strong reason to believe that something was amiss for example, previous JLU associates had reported the existence of a surveillance system and database on innocent users. Perhaps the justification hinges on the evidence one has that more pernicious violations of privacy lie behind the curtains. It would certainly take a powerful battery of arguments to make the case that hacking was justified on the basis that one simply wants to be sure nothing is amiss behind the curtain. Questions about whether to hack or not to hack into a user created organization like the JLU is a kind of moral question for individual users to weigh. From a policy perspective, however, the real question is what the role of the coding authority should be in all of this. The first and most obvious fact is that platform owners and coding authorities like Linden Lab have no business whatsoever collaborating with user organizations that engage in surveillance on their users. It was wrong for Linden Lab to encourage and assist the JLU, and the revolving door between the makers of the Emerald viewer and Linden Lab was also clearly problematic in view of the discovery that Emerald was a black hat technology engaged in violating the privacy of users. Coding authorities like Linden Lab need a zero tolerance policy towards any association with user groups that are engaged in surveillance of other users or that engage in other forms of vigilantism. Such relationships are toxic for many reasons. They give the vigilante groups the impression that they have the blessing of the coding authority, and correlatively feed into the sense that they are the agents of power

and are untouchable. Indeed, they often do become nearly untouchable, as both the JLU and the makers of Emerald (Modular Systems) were successful in getting Linden Lab to shut down a University Campus without due process or even a hint of transparency. But coding authorities need to do more than remain at arms length from organizations that would violate the privacy of their users they also need to make it clear that such violations will not be tolerated and that any operation engaged in such activities will be shut down immediately. As virtual worlds become more and more important to us, and as our lives more and more move online these questions become pressing. Many people now depend upon virtual worlds like Second Life for their business lives, their social lives, and their love lives. It does us no good to have battled for centuries to win the right to privacy in the terrestrial world if we lose it the minute we log in to our jobs or relationships in virtual worlds. Weve made the case that coding authorities must abstain from cozy relationships with vigilante groups that do not respect the privacy of other users, and further that coding authorities need to act decisively against groups that commit these violations, but of course not every such group is known to the coding authorities. What then should be the reaction of a coding authority to a group like The Wrong Hands, which brought the surveillance operations to light? Clearly the kind of reaction favored by Linden Lab banning the associates of The Wrong Hands seems inappropriate. But how do we distinguish surveillance operations like the JLU from a hacktivist intrusion by a group like The Wrong Hands? How can a coding authority reject one set of actions while endorsing the other? Here again there are no easy answers, but clearly there is a difference between a standing organization that systematically engages in surveillance of a broad group of users, and a one-off operation that seeks to expose a specific vigilante organization on the basis of reliable testimony from within the organization. The principle seems to be this: systematic violations of user privacy are not to be tolerated. One-off intrusions designed to expose harm (say an operation like JLU or criminal activity) must be justified and they must be kept at arms length from the coding authority, but if they can be justified then the participants should not be punished. Virtual worlds, like the terrestrial world, need to protect their whistle blowers. Works Cited Alphaville Herald. (2010, September 3). Corsi Mousehold: JLU Are Power Hungry Crazed Nutjobs. Retrieved from http://alphavilleherald.com/2010/03/corsi-mousehold-jlu-are-power-hungrycrazed-nutjobs.html Jenkins, H. and P. Ludlow. (2010, April 14). Choose Your Fictions Well. Retrieved from http://henryjenkins.org/2010/04/choose_your_ficitons_well.html Ludlow, P. (2010, April 9). Watching the Watchers: Power and Politics in Second Life. Retrieved from http://henryjenkins.org/2010/04/watching_the_watchers_power_an.html Mistral, Pixeleen. (2010, May 19). Did Linden Labs Emerald dev coverup lead to Woodbury ban? Retrieved July 12, 2010 from Alphaville Herald: http://alphavilleherald.com/2010/05/did-lindenlab%E2%80%99s-emerald-dev-coverup-lead-to-woodbury-ban.html Mistral, Pixeleen. (2010, May 17). Emeralds devs Modular System datamine tracks 16,740 avatars. Retrieved July 12, 2010 from Alphaville Herald: http://alphavilleherald.com/2010/05/emerald-devsmodular-systems-data-mine-tracks-16740-avatars.html Mistral, Pixeleen. (2010, May 11). Emerald site security broken! Datamining shocks Linden Lab!!!

Retrieved July12, 2010 from Alphaville Herald: http://alphavilleherald.com/2010/05/emerald-sitesecurity-broken-data-mining-shocks-linden-lab.html Mistral, Pixeleen. (2010, April 12). Emerald Viewer: 76000 unique visitors could be wrong. Retrieved July 12, 2010 from Alphaville Herald: http://alphavilleherald.com/2010/04/emerald-viewer76000-unique-users-could-be-wrong.html Mistral, Pixeleen. (2007, June 3). Interview with Woodbury Universitys Edward Clift. Retrieved July 11, 2010 from Alphaville Herald: http://alphavilleherald.com/2007/07/interview-with-5.html Proto, Paladin. (2010, April 20). Linden Labs Woodbury cleanup effort continues. Retrieved July 12, 2010 from Krypton Radio: http://kryptonradio.com/?p=768 Proto, Paladin. (2010, May 17). Modular System Hacked: Woodbury University hacker(s) claim responsibility. Retrieved July 12, 2010 from Krypton Radio: http://kryptonradio.com/?p=854

1 2

Private interview with Burcu S. Bakio lu conducted on 25 May 2010. Private interview with Burcu S. Bakio lu conducted on 24 Deecember 2010.

The Game Behind the Video Game: Business, Regulation, and Society in the Gaming Industry New Brunswick, New Jersey, USA, April 8-9, 2011

Layers of Influence in Multiplayer Gaming


Nadav Lipkin Rutgers University 4 Huntington Street, New Brunswick, NJ, USA ABSTRACT Video game studies literature has not yet been able to theorize multiplayer interaction in a cohesive way. While individual accounts suggest many players resort to barbarism in online play, research often indicates the opposite. To better conceptualize the nature of player agency in multiplayer environments, this essay proposes a diagram describing nine layers of influence on player interaction: the body, the platform, platform developer policies, game developer policies, game code, forums, server groups, individual servers or play spaces, and other players. Parallel to a number of layers are three other layers: interface, clans, and aesthetics and content. The diagram demonstrates that player subjectivity develops through the interaction of virtual, mechanical, social, and aesthetic influence baring sites that, while subject to existing in different ways or not existing at all in some multiplayer scenarios, constitute a more unified system than currently conceptualized. This diagram serves to provide the means of integrating disparate studies on multiplayer activity to reach a more comprehensive understanding of player agency. KEYWORDS: Multiplayer, influence, subjectivity, interactivity, community. INTRODUCTION Breaking News: People Curse on Xbox Live, says the headline of an article on G4tv.com (Bleahy, 2008). The article, mocking the medias surprise that video game players do not always behave in the most civil ways, echoes the sentiment expressed across the internet: video game players online, and especially Xbox Live members, can be really terrible people. The Associated Press (Geranios, 2010) ran a story on Racism and homophobia in gaming: Hate speech corrodes online video game experiences. A satirical video on how to play Halo multiplayer (Fireb0rn) explains, Act like you have a personal vendetta against this person. But Gosh, but I dont, mister! I know, Billy. But that doesnt matter. This is Xbox Live. Act like he pimp-slapped your mom. Blogger Josh Smith (2006) performed an amateur quantitative analysis of the profanity in Halo 2, citing the frequency of individuals with IAS (internet Ahole syndrome). Existing literature already has suggestions as to why online players might collapse into this sort of barbarism despite being well-adjusted in their unmediated lives. Davis (2002) outlines a few reasons for bad behavior on the internet: Anonymity and deindividuation have each been shown to cause aversive behavior in the real world of face-to-face social interactions, and they are even more pronounced in online settings. True identities are difficult to pin down in online social spaces, and reputations do not carry the same weight online that they do in face-to-face interactions. Social presence, the feeling that others are in a social environment with you, is also diminished in online settings. Users may not care if they hurt other users because they have little sense that others are real, little expectation that their bad behavior has consequences for them, and little expectation that they will ever have to interact with the other person in the future (p. 2). McKenna and Greene (2002) echo that notion that deindividuation may increase aggressive behavior by decreasing ones self-awareness (p. 119). Somehow, video game studies literature finds these assertions unsupported in observations. Lin, Sun and Tinn (2003) find that teenager MMO players carefully evaluate risks and benefits to avoid being Lipkin 1

cheated, and lower their expectations of strangers in a manner that we considered very practical (p. 297). Smith (2007) explains, Modern multiplayer games rely heavily on the cooperation between players (n.p.). Pena and Hancock (2006) found that players produced many more positive than negative messages (p. 106). Even blogger Josh Smiths (2006) data on the number of curses per hour in Halo 2 reveal 18.91 curses per hour, the majority of which were not specifically derogatory. One would expect more curses at a Red Sox-Yankees game. How can this difference between qualitative experience and communication theory, which claim players are typically unpleasant, with findings that see players working together more often than splitting apart? This essay argues that differences in civil and sportsmanly behavior and communication between platforms and games on platforms are influenced by different mechanical, digital, social and aesthetic layers available to players in each multiplayer environment. These layers do not operate in a behaviorist paradigm; there is no deterministic cause and effect. These modules, however, serve to encourage or discourage certain actions. The essay will establish the different layers of influence that I propose provide substantive virtual, social, mechanical and, at least partially, aesthetic influences to individual players. This analysis will rely heavily on platform studies, such as implemented in Racing the Beam by Nick Montfort and Ian Bogost (2009), which investigate the relationships between platformsthe hardware and software design of standardized computing systemsand influential creative works that have been produced on those platforms, (p. 2). This essay will attempt to fill the gap in video game studies literature on how the hardware and software of platforms influences, facilitates, or constrains particular forms of computational expression (ibid. p. 3). That is to say that the visible expression of a game, or in the case of this essay, multiplayer behavior mediated through a game, derives fundamentally from certain material and virtual systems, but it does not suggest technological determinism. Parisi (2009) explains ...that we cannot consider only the materiality and design of the game interface, but also the setting in which that game interface is deployed, and the circumstances that surround its deployment (p. 115); platforms exist in cultural space, and players play with them using culturally significant gesturesdifferent players might hold controllers different ways or have different combat tactics in a game, for example. Adding to the concept of platform studies, this essay specifies that platforms are, as Montfort and Bogost (2009) claim, layeredfrom hardware through operating system and into other software layersand they relate to modular components, (p. 3) but they do not exist as platonic ideals; what is now a discrete difference between code and interface could theoretically vanish if some indie programmer were to design a game eliding the two layers so thoroughly that they could not be identified separately. While there may not be a permanent essential character within any particular platform module, necessity dictates that labels be designated; moreover, the labels used in this essay (operating system, game designer, online forum, server, and so on) do not incite enough controversy to warrant complicating. These definitions will be satisfactory until some future game design revolution deconstructs them. The layers in this model bare similarity to the model composed by Lars Knozack (2002) as a method of analysis for computer game criticism (p. 89). He outlines seven layers of potential analysis in games: hardware, program code, functionality, game play, meaning, referentiality, and socio-culture (p. 89), which Montfort (2006) modified to, from bottom-up, platform, game code, game form, interface, and reception and operation. Montfort indicates that referentiality and socio-culture are not subordinate to reception and operation but rather inform the entire model; therefore, surrounding Montforts model is context. The model herein described most closely resembles Montforts with a key exception. In both Knozacks and Montforts essays, the models function as means to analyze computer games (Montfort, 2006, n.p.) and not human interactivity. Therefore, while this model follows a similar course of analysis in some respects, the focus of this model is more separate from individual games and more focused on social influences. LAYERS OF MULTIPLAYER INFLUENCES 1. The Body

Lipkin

Perhaps claiming that the body is the foundational module in such a diagram is unnecessarily complicating; the model could just as easily begin with the highest module being God or the laws of physics. Beginning with the body, however, is indeed a critical initial component to understanding a social phenomenon which is, at least in contemporary video games, inevitably mediated through the player's body. Obvious as it may be, a player who cannot speak will not use the microphone built into their platform hardware. As Parisi (2009) claims, The body is no longer static and disengaged; it is now hailed as a participant in the game text (p. 112). The human body has physical and perceptual limitations which should be considered in discussions of how the social body interacts with other bodies. Even if the body does not play a significant role in a particular examination, it is a required component of any such discussion. 2. Platform Hardware Needless to say, what actions a player is able to take in a multiplayer game are restricted by the mechanical capacities of the platform mediating that game play. As Montfort and Bogost (2009) explain, Sometimes the influence is obvious: a monochrome platform cant display color, for instance, and a videogame console without a keyboard cant accept typed input (p. 3). Therefore, even if people playing each other using a Nintendo DS want to type-chat with each other, the DS as platform hardware does not provide the means for that contact. On the other hand, the fact that it has a microphone and speakers means that, as a platform at least, it is theoretically possible for the DS to provide players a chance to voice-chat with each other, even if no games actually enable that interaction. The gaming platforms hardware is, at its core, the only directly material boundary that constrains player activity. Because play is restricted by the mechanical capacities, peripherals, processing speeds, and interfaces available to that platform, the vast majority of players will experience hardware as the firm limit on the capacity of play, pushing players to use a particular set of communications that are available most efficiently. Of course, the hardware boundary is like the other boundaries in the diagram; it too can be broken. In all fairness, another tier ought to precede the hardware limitations: the laws of physics. Hardware hacking, from turning real guitars into Guitar Hero guitars, toasters into Nintendo Entertainment Systems, and cartridges into hand-held versions of the games on those cartridges, has a strong history with players possessing the skills and ambitions to transform the very physical boundaries of their platforms, modifying them for their particular needs. In theory, a community of hackers could collectively modify their Super Nintendos with microphones and internet connections to provide more avenues for multiplayer interactivity, letting them play Star Fox with each other from different places; however, I have yet to find such an example. Moreover, even in such a case, the vastly dominant group lacks both the technical expertise and the motivation to break this boundary. As such, it is a very rigid one. 3. Platform Developer Policies Beyond hardware limitations on players, other limits could exist in the form of software or distribution restrictions by the platform developers. This is most commonly associated with game content and not significantly with multiplayer interactions, however; Nintendo's policy of restricting the number of games published by individual developers in a year, for example, and its ban on blood, sex, and Satan in games receiving the Nintendo seal of approval are frequently cited examples of platform developers intruding on content decisions. However, approval of what games can be sold is not the major boundary-at least, not for player interactivity, even if it is for game developers and publishers. For players, the major influence of the platform developer is on the software governing the multiplayer interactions; Xbox Live, Playstation Network, or Blizzards Battle.net for example. While game developers have a great deal of room to have different kinds of multiplayer experiences in their games, players on some platforms navigate game developer restrictions through an already restricted

Lipkin

entry put in by the platform developer. As an example, consider that most arcade games cannot prevent players from naming themselves ass or poo, or a variety of different crass expressions. Both Xbox Live and Playstation Network, as well as the new platform for Nintendo's 3DS, forbid players from using certain words, such as curses, rude words for private parts, and racial slurs, in their names (see Ashcraft, 2011). While this current implementation ostensibly has little impact on players interactions in multiplayer gaming, the very possibility for intrusion by the developer into decisions made by the players on their platforms means that the boundary can exist, even if it does not currently in any significant way. The attempts have at least been made: in 2004, Microsoft filed a patent for software designed to censor real-time audio (Bangeman, 2006), and Gabe Newell, the head of developer Valve and the PC platform Steam, suggested, The person who is hated and everyone else leaves, you might want to make that person watch a bunch of ads to help compensate for the negative externalities the[y] create (Purslow, 2011, n.p.). Whether this should be considered a welcomed administrative intrusion or one to be feared cannot yet be debated, but the fact remains that the possibilities remain. 4. Developer Constraints More often than platform developers, game developers play the more significant role in restricting player behavior. Still, the implementation of developer-based boundaries (as opposed to those established in a particular game) tends to be restricted to the level of curse words and racist remarks. In this way, the levels of game developer constraints and platform developer constraints are not especially distinguishablemost physical influence occurs at the level of the game, rather than at the administrative levels above it; yet, while game and platform developers may have similar abilities to shape player interactions, they are differentiated by their relationships to each other. By the nature of the platform, software developers are subordinate to the platform developers, and, as such, what the platform demands the game developer must accommodate and not necessarily the other way around. 5. Game Code As most platforms and developers do not have grand over-arching policies governing the degree to which players can interact across all of the games under their watch, the most significant mechanical influences on player behavior derive from the software capacities within individual games themselves. While platforms may have the mechanical capacities for voice-chat or video-chat, like in the case of the Windows XP OS, not all games enable that interaction. Diablo II, developed by Blizzard, is a computeronly game which ought to be capable of voice-chat; Blizzard's flagship, World of Warcraft is in/famous for its use of voice chat, without which the world would lack the glory of one Leeroy Jenkins. Why, then, does Diablo II lack the ability to have voice-chat when other games running on Blizzard's Battle.net platform can? The answer is not particularly relevant. The important aspect to this distinction is that differences between games in terms of what is and is not permitted within the game's code establish boundaries that guide interactions, either in terms of available communicative channels, such as using chat boxes or voice-chat, or in terms of what abilities players have in relation to each other, such as the ability to squelch players so as not to read messages from them or the ability to kick players out of a game in Left 4 Dead 2. In considering the code limits of a game, one must also consider the proliferation of mods and patches. Both are easily standardized between players and modify game code often in very dramatic ways. In Team Fortress 2, normal players have no ability to kick a player out of the game; however, players who have administrative powers through the Sourcemod program gain that ability and others. Mods and patches as a layer operate similarly to game boundaries, as they both limit players by forcing them to play through the limits of the code; however, mods and patches are subservient to game code in order to operate. They are parasitic, so even when new abilities are created in a mod or patch, they are still bound to the original code. Still, this is a parallel structure, rather than a fully subordinated one.

Lipkin

1-5a: Interface

The interface of a gamethe location of interaction and manipulation between player and technologyis too nebulous a concept to be isolated into a single layer for interpretation. While interface in games most typically refers to the screen or controllers (see Parisi, 2009, and Montfort, 2006), it broadly refers to any location of communication between user and machine, not only including the communication of game content to a players eyes or a players hands to game code. It also includes the players ability to hack hardware and software, physically or virtually modifying content at a more visceral level. While these interactions may not be intended by designers, they certainly exist and can become an integral component to players interactions with a game; consider the proliferation of the Counter Strike mod or home-brew skins, texture packs and graphics patches to Minecraft. The case of the now infamous Playstation 3 jailbreaker George Hotz is only the most recent and most controversial form of this interface with code (see Reisinger, 2011). Interface with company policiesin connection with platform developer and game developer layersis certainly far less common, but it is still relevant to discuss. While interference with games at the level or hardware, software, or presented interface on screen and in-hand can all be accomplished in isolation from any community, interference with company policies requires contact, either welcomed or unwelcomed. Interface with these companies might be as simple as an email or a letter, or it may be as drastic as hacking the company for theft. Exploiting company policy likewise fits into this category, as is the case with a group of hackers who exploited a problem in Xbox Lives Microsoft Points system to generate over a million dollars in stolen money (Warren 2011). This type of interface likely plays little significance in most discussions of games and player interactivity; most players do not interfere with the men behind the curtains. Still, it bears mentioning. SOCIAL LAYERS While the above layers dealt with platforms as they have been traditionally considered, they demonstrate the inability of the current literature on platforms for explaining social behavior independent of the games, their associated hardware, and their peripherals. The layers 6 through 9 focus primarily on social influences on interactivity. Significantly, forums, server groups, and individual servers all have spatial components; they occupy virtual space to which users can refer and have software components that can be manipulated and interacted with in the same ways as the above layers. However, these lower layers are distinct in their social qualities; they typically function in a community, as opposed to the above layers which typically exist in a more static form, independent from community presence. This sense of membership, even if informal, has significant impact on a players sense of belonging and conformity to rules; As group membership becomes increasingly salient, members become highly sensitive to prototypical aspects and use the prototype as a model for evaluating the self and the other group members (McKenna and Green, 2002, p. 122). 6. Forums Forums serve as hubs for communal activity, housing everything from homemade strategy guides to fan fiction to trading posts. Not all forums are created equally, however. Official forums serve as the interface between players and developers, acting as the orthodox line for game-related discussion. The Diablo II official forum on Battle.net, for example, provides players a space to make suggestions, ask help regarding technical problems, and post character build suggestions. However, it is notably moderated by a Blizzard poster, Bashiok, and what the forum moderates is not completely transparent. Presumably, if a player were interested in how to hack the game, use Trojans, or cheat, this information would be blocked. Unofficial forums, such as those on Incgamers.com or Macrumors.com, do not need to tow an official position the same way that Blizzard does with its official forum. As a result, a player looking for a map hack will be able to find information about it in such a forum instead of the official one. While both

Lipkin

types of forums are largely the same, and information that comes from one might quickly end up in another, they are not identical because of their affiliations. Players wishing to contact a developer will not be able to use unofficial forums, and players wishing to subvert the developer will not be able to use the official one. However, for the most part, all forums have a central function; spread information and construct communal consensus. When a player looks for the best location to loot or the fastest place to level up, they will go to the forum where that information is developed, cataloged, and evaluated by other members. The result is the development of a more centralized community of players outside of the boundaries of the games themselves. Regardless of the platform a game is on, forums provide connections between players, even in cases of single player games. One other element of forums is worth considering; because forums do not exist automatically on the same platforms as the games they discuss, there is the potential for a physical distance between the players access to the game and their access to the forum. While in a game on Steam, players can use Steams own commands to access the internet without exiting the game, allowing players to visit forums while still playing. In a more extreme case, a child playing a Gameboy game at summer camp may only be able to have internet access in a single distant location with limited availability. In such a case, physical proximity to the forum is important to how it is used and what questions it will answer. A more typical example would be a console player; players whose consoles are not in the same places as their computers may rely on forums less often for guidance to casual questions because it would require more effort to physically leave the game to consult with the forum. 7. Server Groups The server groups layer refers to an amalgamation of individual game servers under some central authority. The Free Frag Network, for example, consists of a community of Team Fortress 2, Counter Strike, and Day of Defeat players on a number of servers run by a central administration. Each individual server is connected to the others in the same video game, allowing players to quickly switch between servers and track statistics between them. Server groups are significant not because of their consistent virtual locations; the group itself is often decentralized except, sometimes, for a website outside of the game that acts as a homepage. They are significant because they have the capacity to provide the central constitution of rules that all constituent individual servers must obeymuch in the way that platform developer policies on content trump the desires of game developers. 8. Individual Servers/Play Spaces Individual servers here refer to the virtual spaces in which players play the game. The use of the term server, however, should be understood more nebulously; the fact is that no current term fully captures the range of possibilities this layer represents. The play space might be called a server, a game, a match, and so on. However, this layer could be divided into an infinite number of sublayers particular to the game in question. In theory, a server in which a central location branches off into different areasone for trade, one for rest, one for action, and so oncould be considered a server group, as in the above layer. To label and describe every permutation of the idea, however, would be unproductive. This layer, generally, refers to the virtual location of player interaction with what is considered the actual game. In a shooter, this would be where players attack each other; in Minecraft, it would be where players build and destroy together; in Gran Turismo, it would be where they race each other, and so on. It also includes places where trading takes place in-game, such as in trade servers in Team Fortress 2 or the free market portals in Maplestory. In other cases, such as playing Gears of War on Xbox360, this server is the map, hosted by a particular player, when players do combat. In other words, this layer is where play happens. What is most significant about this layer is that it exists in many different forms, and each has significant impact on the reactions players have in them. For MMOs and many computer games, servers have consistency of location; they have stable IP addresses, names, and/or identities to which players can

Lipkin

refer. These server spaces, then, have a concreteness to them, despite lacking a material location in the real world. If a player leaves such a server, they can return to it; even if the servers are hosted by individuals and not on large-scale external servers, if the player who owns the server puts the server back online, it can be found again. Other servers, such as those in many console games such as Street Fighter IV or Halo, host games on the machines of the players that start the games. The servers have no names or identities, and as soon as the hosting player leaves that game, the server ceases to exist and cannot be recovered. MMO public servers combine elements of the consistency of location in computer games and the anonymity of players and inconsistency of the console model. A player who creates a character in the Arcania realm in Maplestory has a consistent location to which to return, in theory, and players in one realm may not be able to migrate to another; however, the shear size of the playing area means that it is not unlike looking for a single person in New York City using nothing but their name. As it relates to the nature of player interactions, this layer is the most important because it establishes the potentiality for relationships based on the duration and proximity of consistent play. Pena and Hancock (2006) explain, Over time, users may also acquire the necessary experience to encode relational (socioemotional) communication using text (p. 96). Accordingly, affectively positive (that is, constructive, joyful, fun) interplayer relationships require time to develop. Logically, the more players can be around each other, the more they can connect to one another and develop their ideal play environment. In the case of a game of Soul Caliber 4 on a console, unique players may be able to play each other multiple times, but unless one adds the other as a friend, they will not likely find each other again, given that matches are created from a pool of all available opponents. The same applies to Halo or Gears of War which select teams from a lobby of all available players. When a player leaves that game, they cannot return to it without blind luck. In contrast, many PC games that run on user-generated servers have consistent locations to which players can return and where the players they previously played with are likely to return. Players acquire to opportunity to learn from each other over a more extended period of play, developing rapport not only between individual players but also to the larger community of players who frequent that server and arguably to the server itself. The result is a server with an affective identity that develops over time. The difference between consistent and inconsistent play environments is the difference between a boxing club and a street fight; a street fight could take place in any alley with any participants, but once that fight ends, it cannot be picked up again with the same participants. Fighting at a boxing club, however, means one can always come back to the club to fighteven if not always with the same exact peopleeventually befriending other fighters who do the same. 9. Players Certainly, it would not be possible to discuss multiplayer interaction without multiple players. While ostensibly, players in games with robot opponents might appear to be playing in similar ways as they would against human beings, these two types of interaction are not automatically identical. For the majority of games in video game history, players played against computer-controlled opponents, developing tactics intended to beat the system: exploits, glitches, cheats, or simply clever tactics to take advantage of their tendencies. Moeller, Espin and Conway (2009) explain, specifically in terms of sports games but arguably true in any game, that there is one guaranteed rule of computer games: whatever a player can do within the environment of a game is allowed and is therefore good (n.p.). Communal debate about whether or not exploiting code is ethical, of course, takes place, but it does not take place within the scope of the game itself. Prior to large scale internet access, friends might argue about an exploit over lunch at school; now, these same debates span the globe on game forums. The stakes, however, for such arguments are low. Jakobsson (2011) describes achievement hunters who typically care more about the accumulated gamerscore than getting all the achievements in any given game. Their approach is to deplete a game of all its time efficient achievements as quickly as possible and then move on. (n.p.). For these players, using cheats and exploits to quickly acquire achievements is personally satisfying and enables them to show off. Jakobsson explains that what seems to oppose hunter activity is the community of players who believe that having a high gamerscore does not equate to any admirable

Lipkin

skills (n.p.). What is at stake here is personal pride and recognition; it does not directly affect any other players. The interaction with human beings can have this achievement and skill based mentality, as would be expected in most competitive spaces. However, players likely take into consideration the differences between playing a human and playing a machine. Bartle (1996) describes what he calls a spade player as having the following motivation: Only in the knowledge that a real person, somewhere, is very upset by what you've just done, yet can themselves do nothing about it, is there any true adrenalin-shooting, juicy fun (n.p.). In other places, this is called being a troll or a griefer. The very existence of trolls and griefers depends on the availability of human players who can be offended and upset. Then again, not all people are trolls. Typical player behavior is more nuanced because, frankly, it is a form of normal face-to-face interaction. While many factors go into the way a player might expect their fellow players to act, such as experience and the servers affect, it is probable that when players do communicate, it tends to be productive. Pena and Hancock (2006) explain, Participants will focus on playing the video game rather than displaying a social orientation when communicating. Thus, participants should produce more task (e.g., 'How can I perform a special move?') than socioemotional communication (e.g., 'Hey, good to see you captain') (p. 96). That same study found that negative comments were less frequent than positive ones (p. 103). While Pena and Hancock (2006) demonstrate a general tendency, they do not consider the importance of hierarchies in their analysis. While this may not be as important in all situations, as a hierarchy is dependent on how much players know and can know about each other, there are many cases when players learn their place in the virtual social order. On many servers, there might be administrators or moderators that police acceptable behaviorplayers who realize that these other players have authority over them will no doubt change behavior to avoid punishment. There may also be non-administrator players who simply play on a server very often and are enculturated into that servers individual culture; a new player on the server might consult one of these regular players for details about the rules on that particular server. Most multiplayer platforms now also have friend lists, and certainly players will interact differently between friends than they would random players they know nothing about. This layer is the last space between the individual player and their actions in the game. 6-9b: Clans Clans (or guilds or corporations) form a particular side-influence for many players in a host of different games. In many cases and most often in MMOs, clans serve as the center for regulating and directing player behavior and activity. In general, clans are community organizations developed outside of a game, typically over the internet, designed to allow players to develop community and have consistent players. It likewise enables players to create and enforce rules in games where they cannot otherwise be enforced. One Diablo II clan recruiter explains, We want to be the best clan out there for people that just want to enjoy the game without being scammed, pvp [player-versus-player], or forced to do unwanted things (PureSaint, 2011, n.p.). Because clans exist in a separate virtual space outside of the game (or are totally decentralized), they cannot be considered a layer above or below a specific layer in the hierarchy. Rather, they encompass the full scope of social layers: they often host their own forums and websites that act as subcultural hubs parallel to more global official and unofficial forums, they establish and run their own server communities, and they operate their own games. Even within games when they may not own or control the entire server space, clan members will monitor, police, and protect each other. One World of Warcraft guild, Wyrd Angles, (2009) explains their operating rules: The Guild is its members. This means taking an active part in the Guild, joining in events, grouping with other Guildies on a fairly regular basis, helping others with quests, doing what you can to help the Guild as a whole (n.p.). This opening line establishes the social nature of the guild, and the page also lists unwelcomed behavior: Don't spam, don't ninja-loot, don't kill steal, don't get caught up in abusive arguments and don't intentionally train (n.p.). Lin, Sun, and Tin (2003) explain one of the other motivations for players to join clans: inexperienced players are happy to find more experienced

Lipkin

comrades to show them the ropes and to back them up against bullies. Membership in a strong clan is thus viewed as a means of self protection (p. 292). In games such as MMOs consisting of large numbers of people at varying levels of skill and ability, clan membership can be a practical importance, not unlike reasons for joining a prison gang for protection. The guild rules for the Intemperance guild (2005) specified disciplinary process: Most infractions will be handled with a warning for a first offense. Repeated offenses will lead to being kicked from the guild. Depending on the severity of the infraction, the council may kick on a first offense at their discretion. The "Time Out Chair" may also be used as a cooling off period if excessive spam is being posted to the guild chat channel or as a 'fun' warning that you are getting close to crossing the line. We will not discriminate. We will put all parties involved in whatever is going on in the Time Out Chair without preference to who may or may not think it right or wrong (n.p.). While the punishment of being expelled from a guild may not seem to carry tremendous weight, it appears to hold a tremendous about of power over members; many guild rule pages list only what actions are unacceptable without even mentioning disciplinary practices. Clan membership is held as an honor, not unlike membership in real-life fraternal orders. In sum, this demonstrates the range of potential locations clans can influence their members; they may influence behavior from separate forums, distinct groups of servers, as well as during play according to clan-specific codes of conduct. As such, it exists as a parallel layer to the social layers. 2-9c: Aesthetics and Content Just as context served as the miasma for Montforts (2006) diagram, the affective role of aesthetics and game content must be placed in a space outside of all of the layers. On purely aesthetic grounds, activity shapes around what content exists and its appearance: websites dedicated to Minecraft use many clean lines and voxel art in reference to the games content; many World of Warcraft guilds feature dragons and other fantasy elements. The complexity comes in defining how significantly content actually changes multiplayer behavior as opposed simply its appearance and aesthetics. Media effects based social science research puts a huge amount of influence in this category (see Anderson and Dill, 2000; Anderson and Bushman, 2001), but most video game scholars dismiss their findings for good reason (see Pena and Hancock, 2006). While this backlash against the influence of content on behavior is a legitimate reaction, it would not be wise to completely separate behavior from content entirely. Researchers who approach games from an interpretive cultural studies perspective found their research on the assumption that content has meaning, and this is certainly the case. Unfortunately, attempting to determine the exact power of content on the individual is not a task that can be accomplished at this time and will need to be addressed in the future. What can be said is that content and aesthetics, along side the spectrum of social layers, contribute to an affective environment which not only changes a players perspective in that multiplayer environment but can also induce a mimetic response to the context. CONCLUSION This diagram of influence is only a starting line. Its purpose is not to reduce video game studies to a formula; rather, it seeks to expand video game studies by increasing its interconnectivity. Rather than examining the mechanics of the platform, the code of the game, the politics of the development process, the social nature of the gaming process, and the visual content of the game as separate areas of inquiry, this diagram seeks to link them together in service of better understanding the nature of the social action that lies at the heart of analysis. It demonstrates that player subjectivity is mediated through a number of potential barriers and affects. Players are not defined exclusively by their environments but they are not autopoetic either. By examining the ways in which players mediate their actions through this maze of pressures and influences, researchers can better understand the nature of player subjectivity generally and extend that into the particular, clearing new ground and enriching the field.

Lipkin

References

Anderson, C.A. and Bushman, B.J. (2001). Effects of violent video games on aggressive behavior, aggressive cognition, aggressive affect, physiological arousal, and prosocial behavior: a metaanalytic review of the scientific literature. Psychological Science, 12.5, 353-359. Anderson, C.A. and Dill, K.E. (2000). Video games and aggressive thoughts, feelings, and behavior in the laboratory and in life. Journal of Personality and Social Psychology, 78.4, 772-790. Ashcraft, B. (2011, March 1). Nintendo 3DS allows racist, sexist and vulgar character names. Kotaku. Retrieved from www.Kotaku.com. Bangeman, E. (2006). Microsoft files for *bleep*ing good patent. Ars Technica. Retrieved from http://arstechnica.com/old/content/2006/05/6757.ars Bartle, R. (1996). Hearts, clubs, diamonds, spades: Players who suit MUDs. Journal of Online Environments, 1.1. Available at http://www.mud.co.uk/richard/hcds.htm. Bleahy. (2008, May 6). Breaking news: people curse on Xbox Live. G4tv.com. Retrieved from http://www.g4tv.com/thefeed/blog/post/685552/breaking-news-people-curse-on-xbox-live/ Davis, J. P. (2002). The experience of bad behavior in online social spaces: A survey of online users. Retrieved from http://research.microsoft.com/scg/papers/Bad%20Behavior%20survey.pdf. Fireb0rn. Bad behavior on Xbox Live [Video file]. Retrieved from http://www.youtube.com/watch?v=PcGuj2q-5Pg Geranios, N.K. (2010, April 17). Racism & homophobia in gaming: hate speech corrodes online video game experience. The Gaea Times. Retrieved from http://tech.gaeatimes.com/index.php/archive/racism-and-homophobia-in-gaming-hate-speechcorrodes-online-video-game-experiences/ Intemperance.org. (2005). Guild rules. Retrieved from http://www.intemperance.org/guildrules.htm Jakobsson, M. (2011). The achievement machine: understanding Xbox 360 achievements. Game Studies, 11.1. Available at http://gamestudies.org/1101/articles/jakobsson Konzack, L. (2002): Computer game criticism: a method for computer game analysis, in CGDC conference proceedings, Frans Mayra (ed.), Tampere University Press 2002, 89-100. Retrieved from http://imv.au.dk/~konzack/tampere2002.pdf Lin, H., Sun, C.T., and Tinn, H.H. (2003). Exploring clan culture: social enclaves and cooperation in online gaming. In proceedings of Level up: Digital Games Research Conference (Utrecht, Holland, Nov 2003). 288-299. McKenna, K.Y.A and Green, A.S. (2002). Virtual group dynamics. Theory, Research, and Practice, 6.1, 116-127. Moeller, R.M., Espin, B., and Conway, S. (2009). Cheesers, pullers and glitchers: the rhetoric of sportsmanship. Game Studies, 9.2. Available at http://gamestudies.org/0902/articles/moeller_esplin_conway Montfort, N (2006). Combat in context. Game Studies, 6.1. Available at http://www.gamestudies.org/0601/articles/montfort Montfort, N. and Bogost, I. (2009). Racing the Beam:The Atari Video Computer System. Cambridge, MA: MIT Press. Parisi, D. (2009). Game interfaces as bodily techniques, in Richard E. Ferdig (Ed). Handbook of Research on Effective Electronic Gaming in Education. 111-126. Pena, J. and Hancock, J.T. (2006). An analysis of socioemotional and task communication in online multiplayer video games. Communication Research, 33.1, 92-109. PureSaint. (2011, February 26). TEEP Clan recruiting. [Web log comment]. Retrieved from http://www.rpgforums.com/forums/diablo-2-clan-hall/126132-teep-clan-recruiting-diablo2clan.html Purslow, M. (2011, February 19). Jerks should be 'charged accordingly' in multiplayer says Valve's Gabe. PC Gamer. Retrieved from http://www.pcgamer.com/2011/02/19/jerks-should-be-chargedaccordingly-in-multiplayer-says-valves-gabe/

Lipkin

10

Reisinger, D. (2011, January 14). Geohot speaks out on PS3 jailbreak legal battle. The Digital Home. Retrieved from http://news.cnet.com/8301-13506_3-20028540-17.html Smith, J. (2006, February 8). Frequency of profanity in Halo 2. [Web log post]. Retrieved from http://tech.gaeatimes.com/index.php/archive/racism-and-homophobia-in-gaming-hate-speechcorrodes-online-video-game-experiences/. Smith, J.H. (2007). Tragedies of the ludic commons understanding cooperation in multiplayer games. Game Studies. Retrieved from http://gamestudies.org/0701/articles/smith. Warren, T (2011, March 9th). Microsoft Points exploit may have cost the company over $1 million. Winrumors. Retrieved from www.winrumors.com. Wright, T., Boria, E. and Breidenbach, P. (2002). Creative player actions in FPS online video games: playing Counter-Strike. Game Studies. Retrieved from http://www.gamestudies.org/0202/wright/. Wyrd Angles. (2009). Guild rules. Retrived from http://www.wyrdangles.net/content/view/5/10/

Lipkin

11

The Game Behind the Video Game: Business, Regulation, and Society in the Gaming Industry New Brunswick, New Jersey, USA, April 8-9, 2011


A Business History of Video Games: Revenue Models from 1980 to Today Joost van Dreunen, Ph.D. Columbia Institute for Tele-Information 3022 Broadway, Uris Hall suite 1A New York, NY 10027, USA ABSTRACT This paper shows that video games are in the midst of a transition from a revenue model centered on packaged software sales to one that charges players small amounts of money for game play. It documents 30 years of video game sales history according to revenue model Arcade, Retail, Subscription, Digital Download and Virtual Goodsand explains key concepts to each of these. KEYWORDS: video games, arcade, retail, digital distribution, subscription, virtual goods 1 INTRODUCTION

To document an industrys history, we may look at different strata. Typically, the technological progress is used as a way to delineate one generation from the next. Financial analysts, for example, will map the video games industry according to the processing power of the various hardware generations (Pachter, 2009). So far, this has been an effective way to chronicle the development of the overall industry, and as a reliable handle in evaluating its various companies. Yet, despite its merits, describing the history of video games this way becomes problematic when technology ceases to be the driving force behind the industrys growth. As the following will show, a substantial part of the industrys recent growth has come not from the traditional, retail-based blockbuster title sales, but from online, mobile and browser-based games that cost less to develop and publish. Consequently, when high production values and mesmerizing graphics are no longer key selling points, we need an alternative point of departure from which to describe this industry. The following discusses the business history of the video games industry and how its underlying revenue models are changing. Drawing on financial data, virtual goods sales, and primary research reports from market researchers and company information, it makes the case that other entertainment and media industries may follow its example. 2 RELEVANCE

The underlying motivation of this paper originates in the video game industrys growth, both in comparison to similar entertainment markets and an economic recession. Figure 1 shows the per capita spending (in 2010 dollars) for the music, movie and video games industry in the United States from 1980 to 2011.

$70 $60 $50 $40 $30 $20 $10 $0


'80 '81 '82 '83 '84 '85 '86 '87 '88 '89 '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11E

Video Game Retail

Music Retail

Box Office

Figure 1: Per Capita Spending on Video Games, Music and Movie Tickets in the United States1 The obvious difference in year-to-year sales for the video games business can, at least in part, be explained by examining how money is made. If we segment the total industrys income by revenue model, it becomes clear that its growth, particularly in recent years, originates in the popularization of several new revenue models, namely digital download, subscription and virtual goods. Combined, these three models have grown from representing 1% of annual video game sales in 2000 to 23% in 2010. Expressing the amount of money a single person spends each year on video games according to revenue, produces the Figure 2.

Numbers shown for video game industry are retail sales only, and include both hardware and software sales. Data for the period between 1980 and 1993 taken from Vogel (2001), chapter 10, pp. 251 263. Data for period between 1994 and 2003 from Noam (2009). Data for the period starting in 2004 and ending in 2011 are originally from Pachter (2009). Data for movie revenues for period between 1995 and 2010 from Noam (2009), table 6.5. Data for movie revenues for period between 1995 and 2010 from Nash Information Services (2011). Data for music retail for period 1980 to 1990 from Noam (2009). Data for music retail for period between 1990 and 2001 from Ziemann (2002). Data for music retail for the period from 2001 to 2010 from DeGusta (2011). Numbers for 2011 (11E) are estimates. All missing values averaged between years. Population data from U.S. Census (www.census.gov/popest). Inflation calculated using Consumer Price Index figures from Bureau of Labor Statistics (http://data.bls.gov/cgi-bin/surveymost?cu)

van Dreunen

Page 2 of 11

$90 $80 $70 $60 $50 $40 $30 $20 $10 $0


'80 '81 '82 '83 '84 '85 '86 '87 '88 '89 '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11E

Arcade

Retail

Subscription

Digital Download

Virtual Goods

Figure 2: Per Capita Spending on Video Games in the United States by Revenue Model2 The traditional ways of monetizing interactive entertainment software, through retail and arcade, in turn show a decline. This warrants investigation. 3 A HISTORY OF VIDEO GAME REVENUE

So far, the American video games industry has gone through five different revenue models. They are Arcade, Retail, Digital Distribution, Subscription and Virtual Goods. Before we dive into each of them, two observations demand attention. First, the different models are by no means mutually exclusive, and often alongside each other. For example, people will spend money both at the video game arcade and on home consoles. Similarly, buying a game for a mobile phone can coexist perfectly with a subscription to an online multiplayer game. Second, a key differentiating factor here is not the shift from one model to the next, but rather their popularization as a critical area of growth for the overall industry. In some instances the technology to, for instance, download games directly to a console, existed long before consumers began to do so en masse. In order for a particular revenue model to become a reliable source of income, a large enough audience base must adopt it. Conversely, by ignoring one, companies risk missing out on critical funds. Advertising is not part of this business history, because it is an indirect revenue stream. All others here discussed involve direct consumer payment. And it is the relationship with the end-user that characterizes the various revenue models.

Data sources for Figure 2 (in alphabetical order): Bagga (2011), Noam (2009), Pachter (2009), Play Meter (2011) and van Dreunen (2010). Population data from U.S. Census (www.census.gov/popest). Inflation calculated using Consumer Price Index figures from Bureau of Labor Statistics (http://data.bls.gov/cgibin/surveymost?cu). All numbers are in 2010 dollars.

van Dreunen

Page 3 of 11

3.1 ArcadeCoin-Ops Charging people a quarter at a time was the first revenue model for the video games industry. In Entertainment Industry Economics, Hal Vogel pinpoints the origin of this pay-to-play model in the late 1880s, with the nickel-in-the-slot machines in the gambling halls of San Francisco. (257) Fifty years later, the emergence of the amusement-only machines provided the blueprint for pinball machines and video game arcades. (258) The coin-op model is deceivingly complex. For players, the quarter per play proposition is simple enough. But the underlying mechanics that need to both reward players and entice them to continue playing, require complex algorithms to monitor, incentivize, and ultimately exploit the players, according to Ely (2009). He explains: The goal was to ensure that a fixed percentage, say the top 5% of all scores would win a free game. The score level that would implement this varies with the machine, location, and time. The algorithm would compute a histogram of scores and set the replay threshold at the empirical cutoff of 5%. Later designs would allow the threshold to rise quickly to combat the wizard-goes-to-the-cinema problem. The WGTTC problem is where a machine has adjusted down to a low replay score because it is mostly played by novices. Then anytime an above average player gets on the machine, hes getting free games all day long. Suffice to say that as a business, both pinball and video game arcade, has a relatively high barrier to entry. According to Vogel, game design can cost up to $1 million per model. In addition, it also requires expertise in other areas, such as manufacturing and assembly, distribution, and maintenance. (260) The heavy machines have to be placed, emptied regularly and kept in working condition, which makes this a labor-intensive business. As a result, the remaining strongholds are venues that can build economies of scale, housing many machines at a single location, rather than having lonely coin-ops scattered geographically. The arcade business model still exists today. But margins have shrunk substantially. The average weekly gross revenue for a single video game machine declined from $129 in 2005 to $69 in 2009. (Play Meter, 2011) Unsurprisingly, year-over-year investments in new machines have also declined by 22% during the same period. In 2010, the Arcade revenue model generated $1.1 billion in sales, down from a peak in 1995 of $6.2 billion (in 2010 dollars). 3.2 Traditional Retail Historically, the video games sector has made the bulk of its revenues through brick-andmortar retail. There are several key characteristics to this model. First, the retail business is largely hardware-driven. Already during its early years, console manufacturers figured out that offering more advanced hardware capable of providing a more sophisticated experience was a strong motivator for consumers. But building, distributing and bringing cutting edge entertainment is expensive. Consequently, the retail model for entertainment software is that of a razor-blade model, where manufacturers accept a loss on the sale of the hardware in the hopes of making a profit by selling software. When Microsoft first launched its Xbox console in 2001 it accepted an estimated $250 loss per unit. Retailers like Wal-Mart, Target and Amazon are examples of outlets that sell video games. In addition, a specialty retailer like GameStop focuses exclusively on selling video game hardware, software and related third-party products. Generally speaking, a retailer demands 20% of the sales price. Table 1 shows the breakdown of margins on different entertainment software sales.

van Dreunen

Page 4 of 11

Retail Price Retailer Take Wholesale Price to Publisher Cost of Goods Sold (COGS) Manufacturing/Packaging Hardware Royalty Fee Licensed Content Royalties Development Costs Total COGS Low High Average Publisher Gross Profit Margin High Low Average Third-Party Distributor

Table 1: Sample Gross Margin Calculation3 PC CDCurrent Gen Next Gen ROM Console Console DVD DVD $49.99 $39.99 $59.99 $10.00 $8.00 $12.00 $39.99 $31.00 $47.99

DS Game Card $34.99 $7.00 $27.99

PSP UMD $39.99 $8.00 $31.99

$2.00 $0.00 $0 - $8.00 $1.00 $7.00

$2.00 $6.00 $0 - $6.00 $1.00 -$5.00

$2.00 $10.00 $0 - $9.50 $1.00 $10.00

$3.00 $4.00 $0 - $6.00 $1.00 $3.00

$2.00 $5.00 $0 - $6.00 $1.00 $4.00

$3.00 $17.00 $10.00

$9.00 $19.00 $14.00

$13.00 $31.50 $22.25

$8.00 $16.00 $12.00

$8.00 $17.00 $12.50

92% 57% 75% $0 - $4.00

72% 41% 56% $0 - $4.00

73% 34% 54% $0 - $4.00

71% 43% 57% $0 - $4.00

75% 47% 61% $0 - $4.00

In this traditional model, the retailers control the value chain because they interact directly with the end customer. This has allowed them great influence on the overall distribution of revenues. For one, a retailer may refuse to carry certain titles that are rated M, for Mature, or A, for Adult. The high degree of concentration in this segment of the value chainwith GameStop, Wal-Mart, Target and Best Buy holding a combined market share of 76% of total retail salesgives retailers a lot of leverage over game publishers. (Pachter, 2009) Secondly, retailers follow tightly organized inventory cycles, which can be very problematic for developers who often have to commit to strict deadlines, early in the development process. This is further compounded with retailers strict rules regarding in-store placement and how long a title is given premium placement. And, third, retailer GameStop also generates revenue from second-hand sales. Allowing customers to trade their used games in for in-store credit is an effective way to cultivate loyalty and repeat purchases. Much to the chagrin of publishers, however, this practice yields them no income. Whereas game companies receive between 34% and 74% on the sale of a next gen console game, the $2.4 billion on GameStop annual balance sheet goes entirely to the retailer (approximately 26% of annual revenue). (GameStop, 2009)

Pachter, 2009, table 31, page 105.

van Dreunen

Page 5 of 11

Important to the retail-based revenue model is its seasonality. About half of annual sales for interactive entertainment software occurs in November and December (Figure 3).
30%

25%

20%

15%

10%

5%

0% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Packaged Software

Figure 3: Month-to-Month sales for Packaged Entertainment Software4 With a two-month window to generate that much revenue, game companies spend a lot of money and effort on marketing. The inevitable crowding that happens as a result, forces publishers to continuously increase their efforts and, ultimately, the involved costs. The retail model still accounts for the bulk of revenue, about 73% in 2010, down from a peak of 81% just four years earlier. Despite this the retail model is depressed. According to the NPD Group, which tracks U.S. retail video game sales, overall revenue for the industry in January 2011 has dropped 5% compared to a year earlier. (Orland, 2011) Retail generated $17.4 billion in total revenues in 2010.5 3.3 Digital Distribution Selling video games directly to an end-user using online connectivity first showed up only a few years after the invention of the home console. But it was not until the mid 90s that this channel would reach critical mass. (Lowensohn, 2009) Because users download the game straight onto their PC, for example, this model side steps physical retail and its cost structure. Here, we distinguish three platforms: console, PC and mobile. It was not for lack of want by the console industry that digital distribution did not reach critical mass earlier. Early initiatives such as PlayCable (1981) and GameLine (1983) died a premature death because of limited hardware capacity, network difficulties, and the video game market crash in late-1983. (Horowitz, 2004) Even efforts in the mid 90s by well-known


4 5

Copeland, 2008. The total retail sales in the U.S. consists approximately of 60%, or $10.3 billion in software sales, and $6.9 billion in hardware sales. (Pachter, 2009)

van Dreunen

Page 6 of 11

companies like Sega, with its Sega Channel (1994), and Nintendo, with the Japan-only Satellaview (1995) and RandNet (1999), failed to gain traction. The current generation of consoles is the first to generate substantial revenue through digital distribution. The PSN, XBLA and Wiis Virtual Console all feature both old and newly released, digital exclusive titles. And a growing number of boxed games is available for digital download, at retail price. Key drivers behind this change are broadband penetration and initiatives by console manufacturers to lower the barriers for smaller studios to develop and publish on their platform. By offering a large library of titles, even at price points lower than those at retail, game companies are able to reach a wider, more mainstream audience. On the PC side, software developers like Id Software began distributing software directly to consumers over the Internet, in the early 90s. After the company released the first iteration of Doom as shareware in 1993, it had reached 10 million installs by the time it appeared in physical retail stores two years later. (Kushner, 2003) Since the PC was a more open platform compared to the walled garden of the console, it addressed a much larger market. Further fueled by a growing number of households coming online, the appetite for digitally distributed content and connected game play started to gain momentum. With the Internet growing steadily, technologies like HTML and Adobe Flash became more sophisticated and popular. Consequently, web-based game play quickly emerged as a common activity for online audiences. On the wings of this development, a host of design companies reinvented themselves as game companies, such as the now-defunct Gamelab, which drafted the original design documents for one of the most popular casual games Diner Dash. Similarly, PopCaps hit title Bejeweled began as a humble Flash game. Initially developers were able to charge around $20 for a downloadable casual game. But because of its relatively low barriers to entry, a slew of low-cost game development companies flooded the market with cheaply produced imitations of popular titles. Unsurprisingly the bottom fell out of the market. Developers found themselves both under pressure of lower prices to stay competitive and in a weakening position with portals who controlled access to the audience. In short order, a companys ability to compete depended on successfully cutting of costs, which ultimately eroded product quality and triggered commoditization. When Amazon.com acquired casual game developer Reflexive in October 2008, it further lowered the average price point to $9.99. And since then the average price of a casual game has crept further down to about $6.99. As a third platform, mobile gaming remained in its infancy until the mass adoption of the Smartphone. Earlier efforts to publish games on mobile handsets suffered from inertia by telecom carriers, who did not to see themselves as content providers. Furthermore, there were immense difficulties associated with the process of developing for up to 100 different handsets, each with their own technical specifications. Carriers, to the chagrin of mobile content developers, insisted that games were playable across the entire portfolio of supported handsets. This policy effectively strangled the potential for an economy of scale, which is elsewhere so integral to digital media. It was Apples 2007 release of the iPhone that re-invented the market for mobile games by offering both developers and consumers a unified platform. In this vertically integrated business modelin which hardware platform, software and distribution all played nice togetherdevelopers are able to publish games with relative low barriers to entry. Consequently, the market has exploded. By January 2011, approximately 663,957 apps were available across all mobile app stores. Apples App Store led the pack with more than two-thirds of total availability, followed by Android Market with about a quarter. A group of second-tier app stores accounted for the remaining share: Nokias Ovi (2%), Blackberrys App World (2%), Palms App Catalogue (1%) and Windows Marketplace (0.3%). Among these apps, games were by far the most popular category in terms of consumer sales.6

Source: company information and app store websites. All sites visited January 2011.

van Dreunen

Page 7 of 11

The benefits of digital distribution make sound business sense. First, it circumvents piracy, an issue that worries many entertainment companies. Requiring a unique identifier for a user in order to play the game prevents people from playing without paying. Secondly, this counters second-hand sales because, once installed, a game cannot be transferred to another user. Digital distribution also, of course, cuts out the traditional brick-and-mortar retailer, allowing for further vertical integration. Thirdly, this opens an opportunity to up sell an existing customer. Making additional levels available after a user has completed the game is an increasingly popular practice. Releasing additional content, such as new missions and characters, for Red Dead Redemption and Borderlands (e.g. new missions, new characters, yielded Take Two Interactive $34 million in digital sales in late-2010, or about 9% of its fourth quarter revenues. (Gamestop, 2010) With approximately 14% of total sales, digital download is the second largest revenue model for video games in the United States, totaling $3.4 billion. 3.4 Subscription Charging users a monthly fee has been a tried and true monetization model for a variety of entertainment industries. So, too, has the video game sector attempted to persuade people to commit long-term. In particular the market for online role-playing games, or MMOs, has been successful in persuading its customers to pay a monthly fee. The first commercial MMO, Islands of Kesmai, launched in 1984 on CompuServe, and cost $12 per hour. That same year, two Essex University students, Roy Trubshaw and Richard Bartle, released a commercial version of their text-based multi-user dungeon. And Marc Jacobs established the company that would later become Mythic Entertainment, and launched a $40 a month MMO called Aradath from a server in his house. This initial success caught the attention of others and soon the increasing competition set a price decline in motion. Quantum Link, the predecessor to AOL, began charging $10 per hour for its graphics-based online service. And shortly thereafter, General Electrics Information Services division released a competing service called GEnie, costing $6 per hour. Finally, as the Internet started to gain momentum in the early 90s, AOL finally changed from an hourly to a monthly rate, setting the standard for the MMOs that characterize the early days of online game play. Within the first two month of operation, Ultima Online (1997), by Origin System (a subsidiary of Electronic Arts), accumulated 100,000 subscribers, each paying $10 a month. Not longer after, Sony and Microsoft added their horses to the race, and released EverQuest (1999) and Asherons Call (1999), respectively. By 2004, the MMO market was in full swing and, after its launch in November, World of Warcraft (Blizzard Activision) reached an impressive 500,000 subscribers by years end. Six years later, World of Warcraft is still the largest subscription-based MMO in the world, with roughly three million active users in North America, and approximately $500 million in annual revenues for the region. Other titles that have successfully adopted this model include EVE Online (CCP Games), Lineage I and II (NCsoft), Final Fantasy XI (Square Enix/Sony) and Runescape (Jagex, Ltd.) Beyond MMOs, casual game companies also discovered the subscription model. After a race to the bottom due to a flood of cheaply produced products, companies like BigFish and PlayFirst began offering subscriptions to lock users in. Here, the monthly expense hovers around $6 a month. But even at such a low price point, BigFish still managed to generate $130 million in revenues in 2010. (Chang, 2010) Companies focused on younger audiences developed their own subscription-based games, such as Club Penguin (Disney). Charging less, around $5.99 per month, the childrens segment grew explosively around 2005, driving growth in overall subscription revenues. The size of this market is substantial. According to KZero, a virtual worlds market researcher, the ten to

van Dreunen

Page 8 of 11

15 years old age group [] totaled 468 [million] registered users in late 2010, compared to 288 million for people aged 15 to 25 years old. (KZero, 2011) Finally, a next generation of subscription-based game services presented itself in the form of OnLive and GaiKai. This type of service enables multi-platform play and does not require the user to own any of the software. While still in early stages of development, the average monthly fee is expected to be around $10. (Takahashi, 2011) Subscriptions currently represent approximately 5% of total video game expenditure in the United States, or about $1.1 billion. 3.5 Virtual Goods As the fifth and final revenue model, virtual goods are here defined as in-game items or game-related services, such as a virtual currency or item, that enable or enhance game play. The principle of selling items and upgrades incrementally as part of a larger game mechanic is a tried and true practice in the toy industry. Virtual item sales, in fact, in more than one way emulate the mechanics behind trading card games. First, in both cases, once an item is created, such as a new character or weapon, it can be copied ad infinitum at no additional cost. Secondly, beyond the use on game play, both trading card games and virtual items can be collected and traded. Consequently, there exist various product categories, like functional items (e.g. a sword), access (for instance to a certain area in the game), temporary abilities like invincibility, and vanity items, which serve a social function. All of these can be tweaked and modified to find the optimal equilibrium between supply and demand, at relatively little cost to the developer. Within the context of video game revenue models, virtual items present a more gradual purchasing experience. In the retail model customers are asked to shell out over $50 before they can play a game at home. The absence of a high initial cost to the consumer, enables people to see if they like the game, and pay as they go. Different than the average selling price of $56.73 for a boxed game for the PlayStation 3 a virtual item may only cost as little as a penny. (Pachter, 2009) Selling items incrementally has proven especially effective in countries like South Korea, where people do not own computers, and instead play at cybercafs. However, when looking at a customers entire lifetimethe total amount of money spent over the course of game playthe number shoots up to $166.85, on average in the United States. (van Dreunen, 2010) Although the audience base may not be as largein an average free-to-play game only between 1-3% of users actually pay to playthe revenue per customer can be substantially higher. In addition, the production process of virtual items allows regular updates, inventory changes, and even seasonal specials and sales events. An important additional benefit of this model is that purchase patterns are a lot less seasonal. In the retail model, a busy holiday season forces many companies to spend a lot of money on marketing in order to persuade customers to spend their annual holiday budget on their game. The lower average cost of virtual items alleviates the dependence on this seasonal cycle. This model, however, does require a game company to be actively involved with their audience. Currently, many of the developers and publishers that used to build a boxed product, now find themselves with a lot of extra work (up to 60% of the necessary resources) after the game has been launched. This forces companies to reorganize their production schedules and resource allocation. And, this also forces game companies to play the role of retailer, having to deal with payment method preferences, distribution and customer service. For 2010, the North American market for virtual goods was $1.3 billion dollars, approximately 5% of the total video game sales in the United States. (van Dreunen, 2010) 4 INTERACTIVE ENTERTAINMENT ECONOMICS

The video game industrys transition to new revenue modelsSubscription, Digital Distribution and Virtual Goodspresents several important benefits over the packaged software van Dreunen Page 9 of 11

model. For one, it alleviates the traditional dependence on seasonality. The sales of virtual goods, for instance, are much more evenly distributed throughout the year than retail sales. Second, it loosens the grip retailers have on the overall production schedule and distribution timelines. This both improves the margins for existing top-tier publishers, and creates room for small studios to work independently. Third, as Juho Amari points out, a more flexible pricing scheme, such as presented by virtual goods, allows for approaching perfect price discrimination. (2009) Other characteristics affecting the interactive entertainment industrys overall way of making money are likely to emerge, since we are only in the early stages of the transition toward digital revenue. As these new models continue to evolve, further research will prove valuable for the games industry and its peers. 5 SUGGESTIONS & RECOMMENDATIONS

More flexible revenue models benefit both consumers and producers of digital entertainment. Therefore, further research into the fundamentals of the Subscription, Digital Distribution, and Virtual Items models is necessary to understand how they may serve other entertainment industries. 6 REFERENCES Bagga, A. (2011). Multi-Channel Game-As-A-Service II: Ubiquitous Games in the Cloud. ThinkEquity LLC, January 24, 2011. Chang, O. (2010). Casual Games Publisher Big Fish On Track to Make At Least $130M This Year. Forbes.com, November 22, 2010. Copeland, A. (2008). Seasonality and Prepackaged Software Price Indexes. U.S. Department of Commerce, Bureau of Economic Analysis, Office of the Chief Statistician. Ely, J. (2009). The Economics of Pinball. GameStop. (2009). Annual Report. Glukhov, A., Jen, M. (2010). Holiday Sales Strength Buoys Take Two. Brean Murray, Carret & Co. Equity Research, December 17, 2010. Hamari, J. (2009). Virtual Goods Sales: New Requirements for Business Modelling? (Graduate Thesis in Information Systems Science). University of Jyvskyl, Dept. of Computer Science and Information Systems. Horowitz, K. (2004). The Sega Channel: The First Real Downloadable Content. Sega16.com, December 21, 2004. Kushner, D. (2003). Masters of Doom: How Two Guys Created an Empire and Transformed Pop Culture. Random House Publishing Group. Kzero Worldswide. (2011). Virtual Worlds: 2011 and Beyond: Key Industry Trends and Market Development. Lowensohn, J. (2009). A Brief History of Downloadable Console Games. CNET News, June 5, 2009. Live Gamer. (2010). Seasonality 2.0. September, 2010. Nash Information Services. (2011). U.S. Movie Market Summary 1995 to 2011. TheNumbers, com. Noam, E. (2009). Media Ownership and Concentration in America, Oxford University Press. Orland, K. (2011). Toys R Us: Video Games Part of Weakest Holiday Sales Category. Gamasutra, January 6, 2011. Pachter, M. (2009). Money For Nothing: How Ancillary Revenues Can Extend The Console Cycle. Wedbush Morgan, Los Angeles, San Francisco, New York, Boston, Seattle. Play Meter. (2011). State of the Industry Report. February 2011.

van Dreunen

Page 10 of 11

Takahashi, D. (2011) OnLive launches its PlayPack library of games for $9.99 a month, VentureBeat.com, February 1, 2011. Van Dreunen, J. (2010). Virtual Item Sales Research Series: North America. SuperData Research, September 22, 2010. Vogel, H. (2001) Entertainment Industry Economics: a Guide for Financial Analysis. (fifth edition). Cambridge University Press. Ziemann, G. (2002). RIAA's Statistics Don't Add Up to Piracy. Azoz.com, December 11, 2002.

van Dreunen

Page 11 of 11

The Game Behind the Video Game: Business, Regulation, and Society in the Gaming Industry New Brunswick, New Jersey, USA, April 8-9, 2011

The 'active' video game consumer: what is at stake in the narratives surrounding the video game prosumer Paolo Ruffino Goldsmiths, University of London New Cross, London, SE14 6NW, UK

ABSTRACT In this paper I will analyse what is at stake in the narratives surrounding the video game prosumer. Although increasing attention is given to the contribution of video game consumers in terms of usergenerated content, software modifications or productions from 'the bottom' (as in the independent gaming scene), there is little awareness of how the discourses describing and justifying these practices are structured. I argue that the assumptions underpinning these narratives of production and consumption are not technologically-determined and do not describe a completely new phenomenon. Instead, they are a redefinition of practices that were already part of the video game industry. I analyse how the producer and consumers of early games and consoles (Spacewar, Magnavox' Odyssey, Computer Space) have been discussed at different times, showing the ways in which these narratives are historically embedded. Other examples, including the game construction sets of the '80s and the 'open engines' of the '90s (id Software's Doom), made different assumptions about the role of the players and developers. From an archaeological perspective (Foucault, 1969), we can highlight the ways in which current discourses of the video game prosumer have been structured, with crucial lessons for the design and advertising of new video games. KEYWORDS: video game culture, prosumer, consumer culture, narratives of production and consumption, media archaeology 1 INTRODUCTION

In the last ten years, the video game industry has become increasingly concerned with consumer involvement in the production of content. Video games such as Little Big Planet (Sony Computer Entertainment, 2008), Second Life (Linden Lab, 2003), The Sims 2 (Electronic Arts, 2004) and The Movies (Lionhead Studios, 2005) are based mostly on user-generated content. Other video games welcome the activity of the players, inviting them to create and share content. In some cases, such as the Xbox Live Arcade, gamers can submit their productions and see them distributed to a wider audience. Some games have met with unexpected success as a result of modifications produced by the gamers. Additionally, new phenomena such as 'independent gaming' seem to be shaping a new video game industry, in which the gamers are finally introduced to the process of production.1 They are no longer consumers, but prosumers. While this is usually described by video game magazines, press agents, and gamers communities as a novelty, or even a revolution, I would like to argue that this is actually a re-definition of practices that were already present among video game consumers. Taking a broader, historical perspective on the role
1

For more information on the Independent Games Festival see http://igf.com


1

Ruffino

played by video game consumers in designing and sharing content, I will look at how discourses surrounding the video game consumer have, in fact, shaped different representations. Discourses around the 'active' video game consumer are based on narratives of production and consumption where the two processes are seen as separate and subsequent. This is historically and culturally placed in a specific age of the medium of the video game. Particularly this appears in 1997-1998 with the release of development kits like Sony's NetYaroze, the release of the open engine of the game Doom and the first edition of the Independent Games Festival. The narratives that describe the production and consumption of games are different in these cases from those that framed earlier cases such as the development kits of the '80s or an 'open' game like Spacewar, developed in 1962. I will further analyse these examples in this paper. Through an examination of specific case studies and examples, I will show the extent to which their emergence and success was shaped by narratives surrounding the video game consumer. I do not intend to argue, however, that the narratives represent a false or misleading description of a culture; instead, I will show how they are at the basis of the structures with which we make sense of video gaming as a cultural phenomenon. Following the methodology of discourse analysis as described by Michel Foucault (1969), I will look at what these discourses and narratives exclude. This process will show how the video game prosumer and, more generally, video game culture is described through structures that frame and enable certain possibilities, while limiting others. I will also provide an original perspective from which to approach the history of this medium and current debates in video game studies, as 'gamers culture' has been brought into the academy without any real consideration being given to the assumptions behind its framing and its historical and cultural contextualization.2 2 METHODOLOGY AND LITERATURE

It is my argument that the narratives of production and consumption in video game culture are based on simple structures, in which characters and semantic values are organized schematically. One such narrative is that of the video game prosumer. This narrative is framed through the schematic opposition of producers and consumers. The 'active' video game consumer, who becomes a producer of content, has already been analysed by several previous authors, including Kcklich (2005) and Sotamaa (2009), both of whom have written on the processes of video game production, approaching the issue from a perspective grounded in the methodologies of political economy. Kcklich (2005) suggests that game modification is a form of precarious labour, demanding awareness from its practitioners. Sotamaa maintains that media consumption and production should not be studied separately, 'as the media practices are becoming increasingly participatory and co-operative' (2009: 4), though he also suggests that 'the increased access to the means of media production does (...) not necessarily equate to increased freedom' (2009: 99). Although acknowledging some issues common to the study of consumer production, both positions are premised on a basic opposition between production and consumption. Here, I turn to Hall's critique of the idea of the prosumer. Hall argues that the distinct roles of producer and consumer, while blurring into the prosumer, are actually reinforced by such a designation: 'production and consumption can be brought together like this in the guise of the prosumer only if they are positioned as having somehow been separate and distinct in the first place which they generally are in narratives of this kind' (Hall 2008: 23).

See, for example, Aphra Kerr (2006), Giddings and Kennedy (2006), T.L. Taylor (2006), Dovey and Kennedy (2006), Alexander Galloway (2007) and Jesper Juul (2009). All these texts describe the video game industry, its market trends, the consumer culture and also cases of 'counter-gaming' (Galloway 2007) where video games are modified by players. Though useful for an understanding of the issues at stake, these texts do not consider how specific narratives of production and consumption initially appeared, or how they are structured. Such descriptions are a part of the object of my research, as statements about video game culture which helped establish the field, and contribute to shape it.
Ruffino 2

I agree with Hall and I also argue that the video game prosumer has not been considered as a result of the processes of its role in structuring the discourses of video game culture. In The Parasite (1980), Serres suggests that a communication model can be re-framed according to a different narrative, in which 'mistakes, wavy lines, confusion, obscurity [and noise are] part of communication' (1980: 12). Similarly, narratives which attempt to describe the subversion of the production models of the video game industry can be (and are) re-framed continuously in the attempt to create a 'grand' narrative (as intended by Lyotard, 1979). However, this attempt is endlessly frustrated by the constant redefinition of these structures, with partially new perspectives appearing in any new statement, as a succession of 'parasites' constantly emerging to supplant their predecessors. In video game magazines, in the interviews with video game players who decide to be producers, in the press releases of consumer games festivals and public events, and in the hospitable reception guaranteed by video game publishers for user-generated content, we can see the narratives of opposition, engagement and independence mixing and contradicting each other. In this complicated succession of views and perspectives, a few elements seem to be constant. Firstly, we can witness the apparent simplicity and linearity of these narratives, often teleological visions of the video game industry, oriented to a progressive opening of the means of production.3 Sometimes they describe the clash between the publishers of mainstream products and the gamers who want to take control of the content.4 In these statements, the linear framing of the narratives conceal contradictory assumptions. One of which, as described earlier, is that a priori separation of producers and consumers. Secondly, some elements are continually excluded by these discourses. In order to preserve clarity and overlook eventual contradictions some points are taken outside of the narratives of prosumption. Here, I follow Foucault's work on discourse analysis, as summarized by Laclau in his piece on Discourse (1995) as an attempt 'to isolate the totalities within which any production of meaning takes place'. Such isolation results in a 'stratum of phenomena', which he calls 'discourse' (1995: 434). His main concern is to understand what constitutes the principle of coherence in a discourse. He first finds it in the episteme, a general outlook, specific to a certain age and which unifies the cultural production of a period. Later, he understands that the episteme does not say much about discursive formations. He suggests then that we should look at the 'regularity of dispersion': the constant interconnection of elements that 'do not obey any underlying or essential principle of structuration' (1995: 435). I will follow this method and look at how the discourses surrounding some previous cases of prosumption in video game culture were framed differently from their current configuration. Firstly, in a historical sense, the very distinction between producer and consumer was far from presumed when the first examples of video games appeared in the '60s, and this precluded the possibility of blurring the two into a new word. As such, discourses about the video game prosumer cannot have been determined by a technological innovation or opening of the means of production. Instead, the framing of the discourses about the video game prosumer appear in a specific period, in the mid-'90s, when the audience of the video game industry begins to shift and change. In this same period, statements originating from the world of open source software began to influence the discourses of video game culture to a significant
In the article New year, new challenges, published on Edge magazine in February 2011 (issue 224, Future Publishing) it looks clear that the age of high-street retailers and high costs of development is oriented to a decline. The article goes even further and questions to which extent the 'indie' world of game development is making it easy to reach a wide public, taking it as an established assumption that nowadays small-scale developers have more opportunities to generate revenue. The teleological vision where the development and distribution of games is going to make it possible for everyone to design and publish a successful video game is taken for granted, as common knowledge. 4 Among the many possible examples, this approach is evident in an article published in the New York Times (15th November 2009). The journalist Joshuah Bearman interviewed the independent game designer Jason Rorher: 'A realization is dawning that games can be much more than what they are now. [...] They have even the potential to be meaningful in deep, fundamental ways. [...] Now anyone can do it[,] which is not how the mainstream video-game industry works'. Available at: (http://www.nytimes.com/2009/11/15/magazine/15videogames-t.html?_r=1)
Ruffino 3
3

degree. It is also a period in which the first historical reconstructions of the history of the medium were published. In these instances, the history of the earliest video games were seen from a perspective based on assumptions about the roles and the distinction of producers and consumers that were of little perceived importance in the discursive formations of that period. Here, I turn to the case of Spacewar (1962), looking closely at the ways in which its development has been narrated in historical reconstructions. I will compare these documents with Ralph Baer's manuscript about the development of Odyssey (1972), the first game console. I will analyse how the narratives of production and consumption were different in these two cases, and how the distinctions set by Ralph Baer's console between software and hardware, producer and consumer, openness and closure, framed further developments in the narratives of production and consumption.

THE 'OPEN ENGINE' AND THE 'BROWN BOX': NARRATIVES OF PRODUCTION AND CONSUMPTION IN THE VIDEO GAME INDUSTRY 3.1 Spacewar, Magnavox' Odyssey and Computer Space

Reconstructions of the history of video games generally agree in situating the 'year one' of digital gaming as 19585, when William Higinbotham, while working at the Brookhaven National Laboratory in Upton, New York, used an oscilloscope to design a game called Tennis for Two for entertaining students visiting the research centre. A few years later, in 1962, at the Massachusetts Institute of Technology, Stephen Russell (also known as Steve 'Slug' Russell in the hacking community) designed Spacewar. The game required two players and was a simulation of a fight between two spaceships. Higinbotham's Tennis for Two was addressed directly to students and visitors, and was not advertised anywhere other than the Brookhaven National Institute. Spacewar, in contrast, was conceived explicitly as software to be shared among researchers with access to one of the first computer models; the PDP-1 (Programmed Data Processor-1). In The History of Spacewar!: the Best Waste of Time in the History of the Universe, an article published in 2009, Matt Barton and Bill Loguidice highlight the fact that: 'far from the secretive and highly competitive world of modern software development, Russell worked in what is now called an "open source" environment, where most code was freely shared and implemented without fear of copyright or patent infringement'. Spacewar was modified by Russell's colleagues at the MIT and in other research centres across the United States. There was no sense of progress towards a final, complete version; rather, the video game's imagined player was an academic researcher who might work at and expand the software. The idea that computers could be domestic tools was still remote and implausible, with Spacewar intended only for gamers with extensive programming skills. Furthermore, the software itself was always open: at any time, it was possible to add or change parts of it. Indeed, it could be said that the possibility to 'close' it, and make the code inaccessible, would have been inconceivable. Russell was influenced by the hacker culture that developed during the early '60s at the MIT and in computer laboratories elsewhere. In this cultural context, the idea of sharing the game's software appeared as the most obvious solution (see Lister et al. 2009: 290). In a similar vein to Russell, Higinbotham would later comment on his own invention by
See Bittanti (1999: 48), Poole (2000: 15) and Anderson, John (1983), "Who really invented the Video Game? There was Bell, there was Edison, there was Fermi. And then there was Higinbotham" in Creative Computing Video and Arcade Games (Spring 1983 issue). J.C. Herz (1997), one of the first texts to describe the video game culture to a large audience, ignores Tennis for Two and places the beginning of the medium of the video game with the invention of Spacewar. Kent (2002: 18) while acknowledging Tennis for Two as the first video game, claims it was an isolated case which did not influence either Stephen Russell or Ralph Baer, who should be regarded as the 'fathers' of digital gaming.
Ruffino 4
5

saying that it looked so obvious to him that he had never thought about patenting it (Bittanti 1999: 50). For later commentators and analysts, however, the reason why these solutions were obvious for them, is often far from clear. It is my argument that this drastic change in the 'obviousness' of the processes of video game development and distribution can be attributed to a shift in the discourses about video games, with an influx of statements and assumptions that had originated elsewhere. As televisions became a widel distributed domestic technology, external devices such as tape recorders began to appear on the market. These were sold as pieces of hardware; 'boxes' that could be connected to the TV set. The product on sale was the device, and not the content, which could be sold separately. I propose that video games, in the process that was about to be initiated by Ralph Baer and his Odyssey game console, occupied a place in a structure originally created for other kind of entertainment devices. In 1965, Sony, Ampex and RCA released video tape recorders for domestic use, while Philips developed its own video cassette format and a specific recorder in 1970 (the N1500, also known as VCR). Philips' American division (Magnavox) released the game console Odyssey in 1972, by which time, as a result of the emergence of the video recorder, the notion that hardware had to be connected to the television set was already firmly established. Ralph Baer reconsidered the possibilities of television devices and Philips/Magnavox applied a model that was already framing their products.6 He applied an existing business model to his product, framing it as something similar to the existing television peripherals. Video games became known to the general public mostly thanks to Ralph Baer and other inventors with an approach to the gaming software that was very different from that of Russell. Ralph H. Baer, an American engineer and inventor, started working on what he called the 'brown box' in 1966. The 'brown box' was a technology for playing games by using a TV set. The product was sold on the market from March 1972 and was named Odyssey. This was the first product for the video game home market. At roughly the same time, inventor Nolan Bushnell had the opportunity to play Spacewar on a PDP-1. He decided to work on a new version of the game that could be commercialized. Reaching an agreement with Nutting Associates, he produced Computer Space, which was presented as a coin operated machine for public spaces, such as pubs, restaurants and malls. It quickly became popular and appeared in public places all over the United States. Ralph Baer and Nolan Bushnell reshaped the vision of video game software. In both cases, the software was embedded in the hardware. Indeed, both Magnavox and Nutting Associates were not selling video game software, so much as machines with video game software installed inside and inseparable from it. This distinction is crucial. Russell first conceived Spacewar as an application of the PDP-1. Ralph Baer and Nolan Bushnell, in turn, worked on machines which supported dedicated software. Ralph Baer referred to his new invention as 'the brown box' in all his personal documents, imaging his invention literally as a box; a closed object which would contain the hardware necessary to run a limited number of games. Stephen Russell's Spacewar, on the other hand, was continually modified to the extent that to attribute the authorship of Spacewar to one single person can only be described as a simplification, made for the sake of providing a straightforward historical narrative. In the book Trigger Happy, Steven Poole offered a full credits list, but even this cannot account for the innumerable modifications that appeared in the following years (2000: 16). There are several issues involved in the passage from Stephen Russell's game to Ralph Baer and Nolan Bushnell's vision of game software. Baer and Bushnell introduced discourses and visions from other domestic industries, envisioning the user as a consumer. In presentations and the documentation of their invention, there are explicit references to the television and pinball industries; a huge departure from Stephen Russell's vision of the user as a potential contributor.

The fact that Ralph Baer thought about his game console mostly as a device for televisions is confirmed by the 'declaration of intents' that introduces his documents for the design of the 'brown box': 'the purpose of the invention is to provide a large variety of low-cost data entry devices which can be used by an operator to communicate with a monochrome or color TV set of standard, commercial unmodified type' (Baer 1966: 1).
Ruffino 5

Further to the commercial interests involved in Ralph Baer and Nolan Bushnell's projects, there is an additional separation which marks them apart from the work of Stephen Russell; a distinction which was carried forward in all following discourses about gamers' games. This difference can be best understood through recourse to the metaphors introduced by Eric S. Raymond (1999) in his discussion of software engineering. He criticized Frederick Brooks' vision of software as a 'cathedral': a work designed by an engineer, the use of which is authorized only after completion, 'carefully crafted by individual wizards or small bands of mages working in splendid isolation, with no beta to be released before its time' (Raymond 1999: 29). There is a certain secrecy about the engineer's plan, one that resembles a form of magic or esotericism. This vision, inspired directly by the theories on software engineering of the '70s and '80s, is criticized by Raymond, who, borrowing from the rhetoric of the open source movement of the '90s, states that software can be engineered instead as a 'bazaar'. According to Raymond, it is better to release software as often and early as possible; allowing users to contribute by finding errors and improving the software, in a way that resembles a disorganized but efficient bazaar.7 Similarly, Russell's game Spacewar was a bazaar, with an unknown, theoretically limitless, number of contributors. Ralph Baer and Nolan Bushnell designed two cathedrals instead, and, in doing so, invented the video game industry and the concept of the video game consumer as a user who could only 'consume' the product. In a cathedral, the visitors face an entrance and an exit; the architecture is complete, self-sufficient, and not to be modified. It is what Baer and Bushnell aimed to achieve with their 'boxes': closed environments that could be accessible only for a specific purpose (to play the game). In a bazaar, however, there is no entrance or exit, and the space can be expanded or closed with no restrictions, as the limits are continually re-defined. When Stephen Russell offered Spacewar to the community of researchers who had access to PDP-1, he applied a 'bazaar' vision of game software, and did not include in his work any conditions on how to limit or close it. I will now look at an historical reconstruction of Russell's story: Hackers: Heroes of the Computer Revolution (1984) by Stephen Levy. Here, we can see how discourses 'act both to constrain and enable what we can know', as McHoul and Grace (1993: 37) write in their discussion of Foucault's methodology. Levy presents a narrative where Russell acts as one of the heroes of the 'computer revolution', despite the fact that such a hero/villain opposition is unnecessary to the story, relying on motivations wholly dissimilar to those that, in reality, moved Russell and his colleagues to action. We cannot reasonably argue what their intentions may have been, but we do know that the discourses that frame video games now are very different from those of 1984, when Levy wrote his piece. Steven Levy includes Stephen Russell in the history of the most famous hackers in the history of computing, with a narration that tends to mythologise the work by Russell and his friends at the MIT. In this book, the hacking culture is defined as inspired by a 'hacker ethic', where authorities are not to be trusted and the only reigning principles are freedom of access (to the computers) and freedom of use. Spacewar is described as a concrete example of this ethic, and the non-commercialization of the game is explained as an example of the opposition between freedom to share and modify software, on the one hand, and, on the other, the restricted access required for commercial exploitation. According to Levy, Russell considered the possibility of commercializing Spacewar only when it was already too late: 'at one point the thought crossed Slug Russell's mind that maybe someone should be making money from this, but by then there were already dozens of copies circulating' (1984: 65). In Levy's view, monetizing one's work is the normal condition; the way it should be. From this perspective Russell's behaviour of not selling the game appears as an ethical decision. The fact that Russell and his colleagues at the MIT failed to even consider this option highlights a difference between their vision and that of those who, like Levy, later reconstructed their story. Russell's decision was in accordance to that which underpinned the discourses on computer programming at that point in time. In Levy's successive description, it appears as an ethical and near-heroic choice. Between
For an analysis of the theories of Brooks and Raymond see Frabetti F. (2009), Technology Made Legible: A Cultural Study of Software as a Form of Writing in the Theories and Practices of Software Engineering, PhD thesis, Goldsmiths, University of London.
Ruffino 6
7

Russell and his commentator, Levy, lies the emergence of the video game market. It was only after Baer and Bushnell invented the video game consumer that the story of Russell, and Higinbotham before him, could be considered as such, and re-presented by Levy. It is interesting, for example, to notice how both Russell and Higinbotham report having failed to consider their software as a potential commercial product. The market had not yet been established and, as such, the notion of video game consumers would have been inconceivable. Thus, rhetoric which questions the emergence of Spacewar, Tennis for Two or any other early example of video game as a missed opportunity for the opening and exploitation of a new market is misjudging the field. If we look at these games from an archaeological perspective, we can see how the conditions established by the discourses in their cultural frame made it incapable to address the production and consumption of a 'video game' (which was not even named as such, as a consequence of still not being a defined object). This is also evident when looking at other contemporary articles about Spacewar. In Jeffrey Fleming's Down the Hyper-Spatial Tube: Spacewar and the Birth of Digital Game Culture, one of the authors of Spacewar, J.G. Graetz, remembers the days when the game was in (permanent) development. When asked about his feelings when a similar game was released, fifteen years later, as a coin-operated machine by a video game company (Space Wars by Larry Rosenthal, 1977), he acknowledges that the idea of copyrighting Spacewar crossed their minds, but they did not consider the full implications: There was a very brief discussion, probably less than a minute, about finding some way to copyright Spacewar, but there were two things; one, nobody knew if it was copyrightable, two, it wouldnt make any money anyway because the game platform was $120,000. [...] We were just having fun. There was no inkling that computers would develop the way they would. [] Nobody knew what programming was. It was something you did to make a computer do things but it had no existence apart from the computer. [] The word software didnt come into existence until just about the time that we got Spacewar done. In fact, the first use of the word in a DEC catalog spelled it wrong. Even after it had a name, nobody knew what it was (2007: 4). However, this level of attention towards the commercialization of the game could not appear in a previous article published in Rolling Stone magazine, in December 1972. Writer Stewart Brand describes the first 'Spacewar Olympics'; a tournament played among the engineers who had access at the Stanford's Artificial Intelligence Laboratory in Palo Alto, California. In this article, Spacewar is a prophetic appearance of the world to come, one where computer will be used not just for work but also for entertainment purposes. While glorifying the game and its makers, this narrative lacked the historical knowledge needed to discuss the possibilities of its commercialization. There could be no such possibility; as a result of Brand's context and circumstance, such concerns were external to the his frame of what was considered possible. Indeed, the final paragraph appendix of the article explains how to 'make your own Spacewar', extending the spirit of contribution to a never-ending project. In 1972, when the Rolling Stone article first appeared, the emphasis was on the liberating effects of computers for the masses. The first few lines made this clear: 'ready or not, computers are coming to the people. That's good news, maybe the best since psychedelics' (1972). J.G. Graetz's comments about the lack of copyrighting for Spacewar also implicitly assumes that the way Spacewar was designed prevented its commercialization, as the required hardware was unavailable on a large scale. The separation between the two, between software and hardware, and the consequences of this separation were not considered (in fact, Graetz underlines how the very word 'software' had not even been coined at the point when they were designing the game). The same article echoes the story of Nolan Bushnell and his Computer Space coin-operated video game for public spaces: 'working out of his home, Bushnell struggled to make the game work on a Data General 1600 minicomputer. Unable to get the economics into the black, Bushnell realized that reproducing Spacewar in hardware, rather than software, was the answer' (Fleming 2007: 4). Bushnell managed to commercialize a video game (actually a clone of Spacewar) only by considering the distinction between the video game software and hardware. From the 'open engine' designed by Russell and his colleagues,
Ruffino 7

the video game as a medium had to become a 'box' a piece of hardware in order to be consumed on a large scale. I argue that this distinction is historically embedded and, as such, it shows how the narratives of production and consumption in video game culture, far from being crystallized, have changed and have played a significant role in shaping this medium. 3.2 Other narratives: construction sets and Doom The understanding of video game software either as a locked 'box' or as an open application of preexisting hardware is something that shifts and changes over the history of gamers' games. This is not a linear relationship, as it is, most often, either suspended between the two notions, or combining elements of both. In recent decades, video game software has started to display again features which were characterised as typical of the 'bazaar'. In this way, it has re-invented the player/user along the line of that imagined by Russell. One example of this can be found in the 'construction set' trend of the '80s. Construction sets were designed as games that included a variety of options for personalizing the game, even to the extent of creating a new one. Construction sets date back to 1983, when the newly founded Electronic Arts company released the Pinball Construction Set,8 which is commonly known as the first game-construction set ever released. Emphasizing the opportunity to become a 'game author' was a first step in separating the role of the producer from that of the consumer. This separation was presented as a narrative whereby an original creator/source would release software to the mass market and then, later, possibly, invite players to join them in this position of power and authorship. The various examples of this 'construction set' game type9 fell short of bringing back the approach used by Stephen Russell's Spacewar in terms of openness and incompleteness. They were still 'cathedrals', in Frederick Brooks' terminology, with the framework of available options provided in advance and protected from further modification. However, game construction sets were instrumental in introducing questions of 'ownership' into the discourses about the video game a medium. Though not consciously problematised at this stage, this issue later provided a base for one of the main discursive fields in the narratives of prosumption; introducing the separation between producers and consumers as a potential opposition. The question of who owns the outcome of the player's performance is something that can be only discussed when the roles of player and developer are separated from each other. One example of how this later posed a problem can be found in the release of Doom's game engine in 1997. As soon as the game engine was freely available, the community surrounding the video game began designing modifications of the game at an impressive pace. Single and multi-player modes spread over the internet. Id Software never claimed the ownership or any copyright infringement, as the game engine was intentionally released so as to guarantee a longer life for their product. In fact, Doom continued to be played and shared and the community of players provided contributions that moved far beyond those of the original game. Here, the documents about its release insisted on considering this structuring of the game as a deliberate decision rooted in a hacker culture of freedom and openness. 10 In the case of Doom, the opening of the game engine is an act that subverted the existing processes of production and consumption. It is described as such because the narratives surrounding this process were

Originally designed in 1980 by Bill Budge and his BudgeCo game company, it was later sold to Electronic Arts which took control of its distribution and publishing. By the end of 1983, the Pinball Construction Set was one of the best sellers of the year. 9 Another popular example was the Shoot'em up Construction Set (1987) by Sensible Software, which is still considered one of the most successful example of a video game construction set. Other popular cases from the same period include Boulder Dash Construction Kit (First Star Software, 1986) and the Racing Destruction Set (Electronic Arts, 1985). 10 This is how the story of the game is considered in Kushner, David (2003), Masters of Doom. How two guys created an empire and transformed pop culture, London: Piatkus.
Ruffino 8

those of a growing industry where the two steps were considered distinct and separate. Furthermore, to leave to the players legal control of their inventions was considered a new and original concession. From the release of Doom's engine, the figure of the active game consumer assumed a new characteristic.11 The kind of structures that Stephen Russell had achieved with Spacewar were beginning to reappear, but on a scale far wider than that of the original PDP-1 users. Moreover, this change arrived when the video game industry was booming as both an economic and a cultural force. Open engines were demanded by gamers in each new release, and this forced a change both in distribution and the relationship between customers and game publishers. Meanwhile, the open source movement was becoming increasingly popular, with openness, the availability of the means of production and the opposition against those who limited freedom of access, emerging as issues in the discourses about software. Video game software was no exception. Game engines offered the possibility to transform a 'brown box' into an open project. Open engines were not only a feature or an added value to a video game, but became a just and fair concession from the development team to its audience of gamers. It is important to notice how these discourses established the game developer as an external figure; defining it as an actor potentially opposed to the gamers in a process of distribution and fruition. Magnavox and Nutting Associates were never described as opposed to the gamers, even as they patented the source code of their games, never allowing it to be available. 4 CONCLUSION

From the late '90s, discourses surrounding gamers' games changed. From this point, it was possible to label video games made with a low budget as 'independent'; modifications to an existing video game software as 'modding'; and video games designed without the necessary license as 'homebrew'. Independent gaming established the image of a 'dependent', corporate business, concerned about the unexpected uses of its products. Modding evoked an imagery of a subversive 'Do-It-Yourself' approach to game software. Homebrew recalled the idea of an illegal development of software, often linked to a re-use of old and abandoned hardware in an unauthorized way. As I have argued in this paper, the situation of the video game industry in the mid-'90s was not essentially 'new'. Instead, it was redefining pre-existing, pre-set discourses of production and consumption. Where active participation of gamers was initially described as a novelty, a category of the active, productive, engaged game players began to surface. The degree of hospitality of the 'original' developer towards user generated content became an issue for discussion and concern, as the game developer became a hierarchical and, indeed, superior actor an original source detached from the multitude of users. From this perspective, the producer could either be opposed to the consumers or could host them, and welcome their participation. This perspective of the prosumer phenomenon is useful, not just in extending our understand of the current configuration of the video game market, but also, more generally, in considering the future development of games and consoles. A deep, historically-based understanding of how a culture represents itself and shapes its values and beliefs is not limited to the current scenario of the video game prosumer, but can be applied to other issues in the video game culture. I would like to maintain, in conclusion, that narratives of production and consumption in the video game culture are not determined by technological advancements. Instead, it appears that representations of the video game culture and technological innovation are engaged in a mutual feedback loop. Furthermore, the distinction between producers and consumers is far from inherent to the industry or cultural form, polarized only under specific discursive conditions and from certain perspectives. In this mutual influence, we can better understand the
11

In the same year, in 1997, Sony Computer Entertainment released NetYaroze, a development kit for its Playstation game console addressed to players who wanted to design and share their own games. It was not the first case of a development kit but it was the first to be fully supported by its developer with dedicated contests and competitions, also hosted on the Official Playstation Magazine.
Ruffino 9

complexities of the figure of the video game prosumer, beyond its apparent, surface-level simplicity and schematic representations. 5 BIBLIOGRAPHY

Anderson, John (1983), "Who really invented the Video Game? There was Bell, there was Edison, there was Fermi. And then there was Higinbotham" in Creative Computing Video and Arcade Games (Spring 1983 issue). Available at: http://www.atarimagazines.com/cva/v1n1/inventedgames.php Baer, Ralph H. (1966), Background material conceptual, TV gaming display, document available online at http://www.ralphbaer.com/Notes_page.htm#1 Barton M. and Loguidice B., 10th June 2009, The History of Spacewar!: The Best Waste of Time in the History of the Universe, in Gamasutra: The Art and Business of Making Games, available at

http://www.gamasutra.com/view/feature/4047/the_history_of_spacewar_the_best_.php
Bittanti, Matteo (1999), L'Innovazione Tecnolodulica L'Era dei Videogiochi Simbolici (19581984), Milano: Jackson Libri Brand, S., (1972), Spacewar. Fanatic Life and Symbolic Death Among the Computer Bums, in Rolling Stone magazine, 7th December 1972, available at

http://www.wheels.org/spacewar/stone/rolling_stone.html
Fleming, J. (2007), Down the Hyper-Spatial Tube: Spacewar and the Birth of Digital Game Culture, in Gamasutra. The Art and Business of Making Games, 1st June 2007, available at

http://www.gamasutra.com/view/feature/1433/down_the_hyperspatial_tube_.php
Foucault, Michael (1969), Archaeology of Knowledge, Paris: Editions Gallimard Galloway, Alexander R. (2006), Gaming. Essays on Algorithmic Culture, Minneapolis: University of Minnesota Press Giddings, Seth and Kennedy, Helen W. (2006), Digital Games as New Media, in Rutter, Jason and Bryce, Jo, Understanding Digital Games, London; Thousand Oaks: SAGE Publications, 2006 Hall, Gary (2008), Digitize This Book! The Politics of New Media, or Why We Need Open Access Now, Minneapolis: University of Minnesota Press Juul, Jesper (2009), A Casual Revolution: Reinventing Video Games and Their Players, Cambridge, MA: The MIT Press. Kent, Steven L. (2002), The Ultimate History of Videogames, Prima Life publisher Kerr, A. (2006), The Business and Culture of Digital Games. GameWork/Gameplay, Sage Publications: London. Kklich, Julian (2005), Precarious Playbour: Modders and the Digital Games Industry, Fibreculture Journal 3(5). Available at: http://journalfibreculture.org/issue5/kucklich.htm. Last accessed 5/1/2010

Ruffino

10

Kushner, David (2003), Masters of Doom. How two guys created an empire and transformed pop culture, London: Piatkus Laclau, E. (1995), Discourse, in A Companion to Contemporary Political Philosophy, edited by Goodin R. E. and Pettit P., Oxford-Cambridge: Blackwell Publishers Ltd, 1995 Lister M. et al. (2009), New Media: a Critical Introduction, London & New York: Routledge Levy, Steven (1984), Hackers: Heroes of the Computer Revolution, Garden City, New York: Anchor Press/Doubleday Lyotard, Jean Franois (1979), The Postmodern Condition : A Report on Knowledge, translated by Geoff Bennington and Brian Massumi Minneapolis : University of Minnesota Press, 1984. McHoul, Alec and Grace, Wendy (1993), A Foucault Primer: Discourse, Power and the Subject, Carlton, Victoria, Melbourne University Press Poole, Steven (2000), Trigger Happy: The Inner Life of Videogames, London: Fourth Estate Raymond, Eric S. (1999), The Cathedral and the Bazaar Musings on Linux and Open Source by an Accidental Revolutionary, O'Reilly Media Serres, Michel (1980), Le Parasite, Paris: Grasset. English version, The Parasite, The Johns Hopkins University Press, 1982 Sotamaa, Olli (2009), The Player's Game. Towards Understanding Player Production Among Computer Game Cultures, Academic Dissertation, presented at the Faculty of Social Sciences of the University of Tampere, Finland Taylor, T.L. (2006), Play between worlds: exploring online game culture, Cambridge: The MIT Press

Ruffino

11

The Game Behind the Video Game: Business, Regulation, and Society in the Gaming Industry New Brunswick, New Jersey, USA, April 8-9, 2011

Call of Duties: the arbitration of online game disputes


Ren Reynolds the Virtual Policy Network London, UK. Melissa de Zwart Adelaide Law School The University of Adelaide, Adelaide, Australia. ABSTRACT When conflicts between players and publishers have arisen, online games have been framed as mere objects of contract and intellectual property. This framing has excluded many of the values that are increasingly at stake, values which a growing number of court rulings and statute are re-establishing. In this paper we argue normatively that one should take cognizance of the fact that online games are ostensibly games. We also make the positive proposal that one can draw upon the rich traditions of law and governance that have arisen around sport when looking at online game conflict resolution. We suggest that policy makers and the industry should consider establishing an Online Dispute Arbitration Board modelled legally, constitutionally and functionally along the lines of sports bodies such as the Court of Arbitration of Sport. KEYWORDS: Governance, Game, Law Sport, MMO 1 INTRODUCTION

Online gaming has grown from its origins in the late 1970 as a niche hobby for a handful of people who had access to a DEC-10 mainframe computer (Bartle 2003) to being a global pastime that is becoming a part of everyday culture in many countries. In many respects online gaming is just an everyday activity people get together and play. The unique characteristic is that this form of play requires technological artefacts in the form of computer, consoles or phones and the Internet. However these material characteristics and the socio-commercial environment in which this form of gaming has emerged has meant that online gaming has tended to be constructed (Law & Hassard 1999); (Law 2001) (Cypher & Richardson 2006), outside of the standard frames of play or sport. This is especially the case in the institutional construction of computer gaming generally in the popular media, policy discourse and the law. Computer gaming has been subject to rhetorics of triviality, aggression and addiction among others (Williams 2003). From the legal perspective, which is the focus of much of this paper, the idea of play and the legal-cultural traditions that have developed around games and sport have been largely excluded from the discourse about computer games. Instead, online games have been framed in terms of intellectual property and contract, which has served to exclude competing discourse about what is at stake (Reynolds 2009) particularly where conflicts arise. Scholars (Reynolds 2002) (Humphreys 2008) (Taylor 2002), some online gaming professionals (Koster 2000), and other bodies (Hogben 2008) have long challenged the legitimacy of this view of online games as mere contract and intellectual property and an increasing number of legal actions and court rulings are further problematizing it. In this paper we argue normatively that online games should be viewed as games by both the legal and policy institutions. This we believe gives access to a rich cultural tradition that encompasses a nuanced relationship between practice, law and governance. Further we shall make the positive proposal that policymakers and industry alike should draw upon the structures of sports governance to establish a body to arbitrate in certain cases of conflict resolution.

Reynolds, de Zwart

ONLINE GAMES

Gaming online is a diverse, global, multi-billion dollar generating activity. The types of games played online include: traditional single player games that can be played within a community e.g. crosswords; traditional two-player or multiplayer games e.g. chess or bingo; single player PC and Console games that have a multi-player component e.g. Halo; massively multiplayer online games e.g. World of Warcraft; and social games e.g. Farmville. In addition to being games played over the Internet there are a number of characteristics that are common to many of the games that fall into the classes listed above. The characteristics that are chiefly notable from a governance perspective are the existence of: Accounts / Characters / Avatars; Virtual Items; and Virtual Currency. Not all online games have all of these attributes, and not all the attributes operate in the same way across any two games, however some combination of these tends to exist in all online games. When disputes arise, especially those that threaten legal action - they tend to involve a conflict over one or more of these attributes (Boyd & Green 2006). 2.1 A typology of on-line game conflict Online games, like any group human activity, generate misunderstandings and conflicts of various types between every possible permutation of the actors involved. Furthermore, because online games are games, individuals are often competing with each other and / or working to attain goals. This heightens the possibility of conflict and adds sportsmanship i.e. generating conflict for competitive advantage, to the run of the mill misunderstandings, rule disagreements, griefing etc. that occur within online games. We can characterise the conflicts that arise in on-line games in terms of the actors that it involves: player, group (guild / clan etc.), publisher or third party e.g. regulatory body; and the domain that the conflict (and resolution) occurs within: in-game, in-context or cross boundary. What we mean by in-game is a conflict that exists and is resolved within the context of the game itself e.g. arguments over loot drops within MMOs. By in-context we mean situations where the conflict and / or the resolution sit within the general context of the game i.e. some element of the state of affairs involves things such as actions in official game forums or lobby areas, mandated arbitration and / or sanctions such as a suspension from a game or forum. Lastly, cross boundary conflicts are those that involve actions such as hacking and resolutions such as legal action. 2.2 Publishers default EULA/IP stance In conflicts that involve a publisher of an online game, publishers have traditionally looked to the legal primacy of contract (generically termed End User Licence Agreement (EULA)) and intellectual property (IP) law as the single frame in which to determine the legal relations between themselves and players of the game and thus the source of conflict resolution (de Zwart 2010). Publishers even make reference to this in debates on official forums. 3 CROSS-BOUNDARY CONFLICTS AND REGULATIONS

Various forms of cross-boundary conflict have given rise to a rash of threats of legal action, a number of cases (both criminal and civil), several statutory changes specially addressing online gaming, and a number of other forms of regulatory action. The general trend of these judgment and statute has been to erode the standing of publishers default EULA/IP position. In some cases publishers defence of their EULA has been seriously challenged (Bragg v. Linden Research 2007). While this emerging state of
Reynolds, de Zwart 2

affairs is recognizing more rights in players it is also problematic as it indicates both a greater value, in a general sense, in online games and a greater uncertainty as to their legal status. This may be illustrated by examining a number of incidents and related statutes and regulations from a range of jurisdictions. 3.1 RMT Real Money Transactions Much of the debate about the status of virtual items has centred on so-called Real Money Transactions ((RMT) i.e. users of a game selling virtual items, money or accounts for hard currency). Generally publishers have argued that RMT is in breach of contract, whereas those defending such actions (or seeking to recover their in-game assets) have argued along the lines that they accumulate rights to virtual property through their labour. Publishers act on the basis of the EULA when they ban players for selling virtual items. This is a constant activity in the publishers struggle with so-called gold farmers and is seldom challenged. One of the first actions that was threatened on behalf of users in respect of RMT related bans was the 2001 Gravity Spot threat of a US based class action against Sony and Verant for closing down auctions of virtual items on sites such as eBay. While this seemed not to get anywhere beyond a web page and a few headlines (the web site no longer exists) it did strike the tone taken up by later user complaints and actions. In 2002 Blacksnow Interactive (BSI) initiated an action in California (US) against Mythic Entertainment - publishers of Dark Age of Camelot. This case included the issue of Mythic banning accounts and used the concept of unfair business practices to argue against it. The case fell apart when the individuals behind BSI seemingly disappeared leaving their lawyers unpaid and it has was later alleged that BSI had established the first virtual sweatshop (Dibbell 2003). The situation is rather different in Asia particularly in China and South Korea where judicial guidance and statute have started to define what is permissible in respect of RMT, whatever the EULA might state. In 2006, South Korea introduced the Game Industry Protection Act that was subsequently amended in 2007. In summary, South Korean law makes it legal for an individual (not commercial online exchanges) to sell virtual items derived from games of skill provided that such virtual items were gained through normal play and the earnings were declared and within prescribed income limits. In effect this makes it possible for individual players to sell things earned from games but prevents so-called gold farmers from running their businesses. The Supreme Court of South Korea clarified the law in relation to MMOs in 2010 when a case of RMT trading of Lineage (NCSoft, Korea) in-game currency Aden was brought before it. Here the Supreme Court overturned the lower courts view that MMOs are a game of chance, noting that constant efforts to get as much Aden as possible can also be regarded as a game that requires much time and effort. Furthermore, no evidence was provided to demonstrate that any of the currency had been gained through abnormal play - meaning the defendants were acting within Korean law irrespective of the prevailing EULA. In China, the Ministry of Commerce announced RMT related rules in both 2009 and 2010. The 2009 rule prevented virtual currency being traded for out of game currency, goods or services. The 2010 rules banned minors from games where hard currency can be exchanged for virtual currency; it also required real-names to be used when registering for online games (People's Republic of China 2009). 3.2 Theft This first set of incidents and related cases arises where individuals (or groups) take virtual items from a player account, gain money from a player for virtual items through deception or gain personal and credit card details from individuals through deception centred on virtual items. This set of acts falls within the general categories of hacking, extortion or duping (Hogben 2008).

Reynolds, de Zwart

Typically when many of these cases have been reported in western media they have been headlined player convicted of virtual theft. Despite these headlines, domestic courts have usually dealt with these cases on the basis of laws relating to unauthorised access to a computer, computer fraud or similar. From a player perspective this has been like arresting someone for breaking your front door but not for anything they did in your house. There are however exceptions to this approach in both Asia and, more recently, in Europe. In China in 2003 Li Hongchen brought a case against Beijing Arctic Ice Technology Co. Ltd. This case concerned a player whose account was hacked and virtual items transferred to another account. The publisher refused to provide the details of the other account holder or re-instate the items. Beijing's Chaoyang District People's Court ruled that the publisher had a duty of care to the player and was ordered to restore the items, which the judgment termed the players property. In 2007 a 17 year old was arrested in The Netherlands for stealing virtual items in Habbo Hotel. The person had hacked into accounts and transferred items out of them the plaintiff was charged with crimes related to Computer Intrusion and Destruction (because the items were effectively destroyed from the hacked accounts) and also theft. As the Dutch authorities said they were also charged with theft and receiving stolen virtual goods because they still represent a value [authors translation based on Google translation]. The suspect was not convicted. Also in The Netherlands in 2007 a 13 year-old player of the online game RuneScape (Jagex Games Studio, UK) was kicked and threatened with a knife by a 14 year old and a 15 year old until he transferred virtual items to one of their accounts. In 2008 a Dutch court found the both defendants guilty of robbery under Article 310 of the Dutch Criminal Code, noting that the virtual items qualified as goods under Dutch law. The defendants were sentenced with 180 hours of community service and 'youth detention for four weeks with a probationary period of two years. This year in the UK an individual was arrested and admitted to hacking into the accounts of online game provider Zynga, transferring around 400 billion virtual poker chips into his own account. He then began selling the poker chips on the black market at a price much lower than the cited $12 million face value (Anon 2001), and was convicted of 4 counts of Converting Criminal Property and 1 count of Unauthorized Access (under the English Computer Misuse Act (Great Britian 1990)). At the time of writing the plaintiff awaits sentencing. 3.3 Other cross boundary disputes The cases summarised above all focus on virtual items but while they offer quite different legal views of the nature of virtual items they still put notions of property (or its rejection) at the centre of the rhetorical stage. There is another set of actions, all brought by users, where such notions are either incidental or irrelevant. The most well known of these actions is the so-called Blizzard GLBT incident that occurred in 2006 when player Sara Andrews advertised in in-game open chat for new members of a GLBT friendly guild. Another player reported this to the company and Ms Andrews was sanctioned under Blizzards Terms of Use section on Harassment Sexual Orientation. Following a public outcry (Calleja & et al 2006) and the threat of a lawsuit Blizzard backed down and apologised to Ms Andrews for the incident. In 2009, (Estavillo v Sony Computer Entertainment America 2009) Erik Estavillo, who is a serial litigant against computer game platform providers, brought an action against Sony with respect to Sony banning him from the PlayStation 3 Network (PSN) for use of bad language within the game Resistance. While the serial nature of Mr Estavillos litigation following this action has caused the press to raise questions about his motives, the nature and fact of the cases is very interesting. In this first case, Estavillo made claims on three grounds: restriction of free speech; theft of money in the form of the
Reynolds, de Zwart 4

virtual currency held in the PlayStation Network Wallet Fund (which was confiscated when he was banned from PSN) and inconsistent application of governance and game rules e.g. under 18s regularly playing 17+ games. The judge denied Estavillos claim stating that: Although the Network does include "virtual spaces" such as virtual "homes" and a virtual "mall" that are used by a substantial number of users...these "spaces" serve solely to enrich the entertainment services on Sony's private network. In providing this electronic space that users can voluntarily choose to entertain themselves with, Sony is merely providing a robust commercial product, and is not "performing the full spectrum of municipal powers and [standing] in the shoes of the State.' Lastly in 2010 Stern brought an action against Sony (Stern v Sony Corporation of America 2010) claiming that they should make reasonable modifications to their games to accommodate players with disabilities. In particular, the plaintiff, who had visual and other learning difficulties, required Sony to provide or enable mods which would provide visual and auditory cues. The judge dismissed the claim on the basis that the relevant provisions of the Americans With Disabilities Act (ADA) apply only to physical places or goods or services connected to physical places. The claim based on the fact that inability to fully participate in the games limited the plaintiff's ability to participate in gaming conventions organised by Sony was also dismissed. The case essentially failed on the basis that it was brought under a public accommodation clause of the ADA and online games were held not to fall under this part of the Act. Opinions range on whether the underlying merits of the case would be successful if a slightly different legal approach were adopted (Lastowka 2010a). 4 DISCUSSION OF TRENDS

The events summarised above provide a range of interpretations of what is at stake when a crossborder conflict occurs in respect of an online game. The meaning and value attributed to various aspects of online games vary depending on the nature of the conflict. We would expect this. After all a humble stick can be variously: a weapon, a piece of fuel, an instrument etc. depending on the context in which it is viewed so we would expect a similar plurality with things as complex as online games and the various practices associated with them. What is more problematic are the conflicting and potentially inconsistent views of how they should be treated. This is complicated further by the international nature of online games particularly where people from a range of countries are playing in the same online space. While the authors view games as a social good this could prove to be increasingly difficult for publishers to manage should there be an increase in conflicts brought before the courts which we predict will be the case. For example, the theft and Chinese duty of care cases discussed above, all suggest that virtual items have a value that is not merely that of the publishers intellectual property, these items can be stolen and their value should be protected. A policy maker or regulator might view this state of affairs and conclude that virtual items and especially virtual currency are so akin to e-money that they demand an equal level of protection and due diligence on behalf of the publisher. This would have a direct and no doubt chilling impact on the online game industry, as publishers would have to take a large step closer to being banks. But the situation where virtual credits such as Xbox points (Microsoft) are bought for hard currency and immediately contractually defined as something that a user has almost no rights to seems untenable. It is unlikely that domestic consumer protection authorities would allow this state of affairs to continue indefinitely. To return to our initial assertions, the fact that online games are games, and thus virtual items are game objects adds yet another level of complication. That is, where law and regulation does intersect with online games it must take into account not only the values attributed to virtual items by virtue of the commercial and ludic relation to them that players have, but also the way that they function within the
Reynolds, de Zwart 5

context of a given game. For example while a space ship in the game EvE Online (CCP, Iceland) might have an apparent value of many thousands of euro the rules and practice of the game make it clear that that any ship might be irretrievably blow up or stolen via subterfuge at any time and any form of regulation should protect not put into jeopardy this state of affairs (de Zwart 2009). 4.1 Assumptions and a way forward Given the trends indicated above we suggest that a number of things are likely to occur in the short term: Increased theft of virtual items; Increased third party sales of virtual items and related crimes e.g. credit card fraud; More players will seek legal remedies for publisher acts that they do not agree with; More judgements will be found in favour of players overturning EULA and copyright primacy; and, Policy makers and regulators will increasingly focus on the virtual items and currency probably creating regulations that inadvertently harm at least one sector of the industry, thereby also harming citizens that enjoy participating in that sector. However we also believe that many of the issues that may arise have a close analogy with sport and that there is a rich tradition of sports law and regulation whose underlying principles and possibly analogous governance structures may be applied to online gaming. What we wish to highlight in the following discussion is the role and recognition given to bodies that sit between participants and national laws. 5 THE GOVERNANCE OF SPORT

The relationship between sport and law is complex. Central to this relationship is the fact that when stripped of their context, acts that occur in many sports would be seen as curious at best and illegal at worse. Contact sports, especially the more physical ones such as boxing or rugby, are made up of a series of acts that in any other context would be considered as assault, resulting in injury and sometimes death. However sports people are not regularly arrested for their acts, despite the overwhelming number of witnesses and records of the events (Lastowka 2010b). Rather, the law accommodates the cultural values embodied within sport, but it does so in a complex, dynamic way. Sports by their nature have rules. These are governed by the physical nature of the sport e.g. the shape of the ball; on the ground officials of Figure 1: Governance of Sport some type e.g. umpires and referees; clubs and other bodies; then in the last resort law. Infractions of rules, even those that might otherwise be considered criminal, are regularly left to non-state actors who have the requisite governance and normative power e.g. the power to exercise practical sanctions. When cases do go to court they are judged contextually i.e. a court does not ask did the rugby player hit the other player (as of course they did, they do all the time) but rather did they do so with intent and in a manner that is outside what would be reasonably expected given the context and rules of the sport. In practice, things are more complex than this as there may be disputes regarding jurisdiction. However all of this occurs within an eco-system of governance made up of a set of inter-related actors each of which have relevantly well understood domains of consideration, sanction and hierarchal standing e.g. a governing body can overrule and official, and a court can overrule a governing body.

Reynolds, de Zwart

What is significant for our purposes here is that all parties recognize both the centrality of game derived meaning, value and harms; and, recognize how the limits of those values are negotiated. Whats more, the role of rules in generation of value and the adherence to stated rules in the resolution of conflict are also given primacy. All of this appears to have direct application to online gaming. Further it should be noted that sports in general have constructed normative governance structures that resolve conflicts and apply sanctions as a layer between participants and national law, and that courts recognise the regulatory power of these bodies with understood and negotiated limits. 5.1 Sports Governing Bodies Typically in sports there is a hierarchy of governance: event officials, team, league, national governing body, regional governing body, international governing body, Court of Arbitration of Sport. The purpose of these bodies is to define the rules of a sport, to apply them and to regulate their application. For example a player may face an in-game sanction for breaking a rule, following this a national governing body might suspend that player for a period of time, the player may appeal this through the national governing body or against the national governing body via an international governing body. National law may be involved in cases where the infraction might be considered criminal and / or if the player takes action again a governing body for misapplication of rules (typically brought under breach of contract (James 2010). Appeals against decisions do not simply flow up a governance hierarchy, as each level tends to define its scope of jurisdiction, often leaving those closest to the sport or given event precedence over certain matters. The Court of Arbitration of Sport for example tends not to review in-game decisions of sports officials (James 2010). Historically sports governing bodies were formed by practitioners and / or ex-practitioners of a given sport and were loosely defined associations of peers. Today governing bodies tend to be companies limited by guarantee. Governing bodies tend to have direct contractual relationships with the bodies above and beneath them in the governance hierarchy. Often players either have no written contract (in amateur sport) or a contract of employment with their club (in professional sport) hence no direct contractual relationship with a governing body. In cases of legal dispute courts typically infer contractual relations between players and governing bodies through conduct on both sides (there are some exceptions to this (Hartley 2009). 6 APPLYING SPORTS GOVERNANCE TO ONLINE GAMES

We believe that the lessons of sports governance are applicable to online games as in both cases the governance structures relate to and act upon a state of affairs whose value and meaning are defined by a practice that is outside of the ordinary and would be treated differently by law if it were not for its context. That is, in a sense, a punch thrown in a boxing match is akin to the value an individual places on virtual item held in an online game account: one is-and-is-not just a punch the other is-and-is-not just a pile of data. 6.1 Online Dispute Arbitration Board: ODAB Therefore, we argue that there is a possible role for a governance body in online games. For the purposes of this paper we will title the putative governance body the: Online Dispute Arbitration Board, henceforth ODAB. An alternative arrangement is an Ombudsman (Online Dispute Ombudsman ODO) system that could exist on a publisher-by-publisher basis. The proposal here is not incompatible with this, however we feel that a federated body is simply less costly to implement considering logistics of multiple terms of reference etc.
Reynolds, de Zwart 7

Where in the hierarchy would the body sit? Given that the issues that are starting to arise tend to be between users and publishers of online games, the most useful role for ODAB would be to sit between players, publishers and national / international law. Whats more it seems to us that the issues that arise while having some online game specifics tend to be of a generic nature relating to the application of the publishers own rules. Hence, while we see an argument for a governance layer made up of publisher / game specific ombudsmen (and do not rule this out as an additional option), we believe that ODAB would be of most benefit sitting across a set of online games that accept its governance. What would be its scope of action? Applying the language of the Court of Arbitration of Sport the Mission of ODAB might be defined as: Mission - The ODAB sets in operation Panels which have the task of providing for the resolution by arbitration and/or mediation of disputes arising within the field of online games and social networks in conformity with its Procedural Rules. To this end, the ODAB attends to the constitution of Panels and the smooth running of the proceedings. It places the necessary infrastructure at the disposal of the parties. The responsibilities of such Panels are, inter alia: (1) to resolve the disputes that are referred to them through ordinary arbitration; (2) to resolve through the appeals arbitration procedure disputes concerning the decisions of publishers, associations or other online game / social network related bodies, insofar as the statutes or regulations of the said online game / social network bodies or a specific agreement so provide . It is envisaged the operational scope of ODAB would likely include: Appeals against bans from games / social networks; Exclusion from virtual items be this through: Ban, Removal of items through means outside the rules of the game e.g. hacking, Publisher error e.g. accidental deletion. The scope of ODAB would exclude: Changes to game mechanics e.g. nerfing; User vs User disputes; and, User vs Group e.g. Guild disputes.
Figure 2: Online dispute resolution

Social networks have been included in the definition above as it seems that many of the same issues and disputes and modes of resolution that apply to in-game items, especially those in games based within social networks, apply equally to access to the network itself. What would be its relationship with other actors? ODAB would have contractual relationships with either individual publisher and / or trade bodies. The bodies would fund ODAB through fees related to the number of users they have the system of charges needs to be detailed due to variations in business models e.g. freemium where there is a non-lineal relationship between users and revenue. How would it be constituted? ODAB would be a company limited by guarantee / not for profit legally constituted in the US, EU and Australasia. Who would serve on it? ODAB would be made up of both publisher and players. A constitution would be required to determine what individuals held what roles for what periods etc. Why is this better than the current arrangements? If the assumptions above relating to the number of conflicts over online games that go to court are correct, then the benefits already found in sport governance should accrue to online games, these are:

Reynolds, de Zwart

Cheaper - Arbitration tends to be cheaper than law. Even when quasi-judicial bodies are established these tend to be less costly than legal actions. Further, the existence of an arbitration body does not preclude the option of legal action, as we see in sport. Better - Assuming that ODAB can attract the right mix of individuals it should provide a better process than the courts, at least in the short term. This is because, as with sports, ODAB will have individuals that understand the details and culture of online games, hence will understand the issues at stake for all the actors. Regulatory burden - A governing body that has normative power on players and publishers relieves the burden on states, at least in part, from enacting legislation and creating statutory bodies to deal with the issues that begin to arise in greater numbers from online games. This is particularly important as online games tend to be international and states tend to act first on a heterogeneous national basis and then take some years to come to forms of international consensus all of which is time consuming and costly. Regulatory peril - From a publishers perspective, an independent arbitration board may give states confidence that citizens will be sufficiently protected as customers of online games such that they do not need to pass the legislation and create statutory bodies noted above. The peril for publishers is that any such action runs the risk of having serious intended or unintended consequence on the industry as it is hard to pass laws that capture the nuance and dynamic of individual games. Here again the sports model is apt as states tend to regulate by law the rules of individual sports. 6.1.1 Objections Insufficient volume currently there is a low volume of disputes that are taken to court hence costs of any arbitration body are likely to outweigh any savings from the few disputes it might hear. As we stated in the assumptions above we feel that the number of cases will increase but we concede that it is difficult to know when the best time to set up an arbitration body would be and that it may initially be financially inefficient. It will be griefed almost all systems of online game norming are exploited by some players (Foo & Koivisto 2004), the same will happen with an arbitration board with players simply wanting a cheap way to grief publishers. The scope of the arbitration board must be defined such to filter out griefing, for example a rule would be that publishers dispute systems are fully exhausted before the arbitration body is evoked. There is no publishers association the online game industry has no recognised body hence it is unlikely that publishers will recognize the arbitration board. Initially the arbitration board will have to work directly with publishers and existing associations e.g. some publishers of online games are members of general publishing associations such as UKIP in the UK and ESA in the US. More broadly the lack of self-identification of online publishers is seen as a barrier to the creation of an arbitration body. It is not peer based / lack of player representation it is likely that the body will be made up of publishers only hence will be bias and fail due to lack of credibility. This is a challenge for the body especially as funding is likely to come from publishers. Thus the constitution of the body that publisher agreed to must ensure representation of players either through other bodies or non-publisher individuals. 7 CONCLUSION

The rise in popularity of online games and the social function that they and in-game virtual objects are accruing is changing the values and meaning given to those objects. Courts are recognizing this in
Reynolds, de Zwart 9

decisions that overturn the EULA / IP basis of publishers traditional arguments. A governance body that sits between publishers and national law seems to us to provide the most appropriate route to justice in the cases of dispute that we predict will increase over the next few years especially given that the alternative is likely to be heterogeneous national laws and regulation. We thus call upon the online games industry and other interested parties to consider the formation of an online dispute arbitration board. REFERENCES: Anon, 2001. A COMPUTER hacker from Paignton has admitted stealing $12million worth of poker chips from an American gaming company. Herald Express. Available at: http://www.thisissouthdevon.co.uk/news/HACKER-ADMITS-STEALING-12m-POKERCHIPS/article-3170994-detail/article.html [Accessed March 13, 2011]. Bartle, R., 2003. Designing Virtual Worlds, New Riders. Boyd, S.G. & Green, B. eds., 2006. Business & Legal Primer for Game Development, Charles River Media, Inc. Rockland, MA, USA. Bragg v. Linden Research, 2007. Inc 487 F. Supp 2d 593 (E.D. Penn. 2007), Calleja, G. & et al, 2006. Open Letter to Blizzard Entertainment. Available at: http://terranova.blogs.com/terra_nova/2006/02/open_letter_to_.html [Accessed March 21, 2011]. Cypher, M. & Richardson, I., 2006. An actor-network approach to games and virtual environments. In Proceedings of the 2006 international conference on Game research and development. Perth, Australia: Murdoch University, pp. 254-259. Dibbell, J., 2003. The unreal estate boom. Wired Magazine, 11, p.01. Estavillo v Sony Computer Entertainment America, 2009. 2009 WL 3072887 (N.D. Cal. Sept 22, 2009), Foo, C. & Koivisto, E., 2004. Grief player motivations. In Proceedings of the Other Players Conference. Great Britian, 1990. Computer Misuse Act 1990, London: The Stationary Office. Available at: http://www.legislation.gov.uk/ukpga/1990/18/contents [Accessed March 23, 2011]. Hartley, H., 2009. Sport, Physical Recreation and the Law New Ed., Routledge. Hogben, G. ed., 2008. Position Paper: Online Games and Virtual Worlds. Available at: http://www.enisa.europa.eu/act/it/oar/massively-multiplayer-online-games-and-social-andcorporate-virtual-worlds/security-and-privacy-in-virtual-worlds-and-gaming. Humphreys, S., 2008. Ruling the virtual world. European Journal of Cultural Studies, 11(2), p.149. James, M., 2010. Sports Law, Palgrave Macmillan. Koster, R., 2000. Declaring the Rights of Players. Available at: http://www.raphkoster.com/gaming/playerrights.shtml [Accessed March 13, 2011]. Lastowka, G., 2010a. Is Everquest II a Place of Public Accomodation? Available at: http://terranova.blogs.com/terra_nova/2010/03/is-everquest-a-place-of-public-accomodation.html [Accessed March 21, 2011]. Lastowka, G., 2010b. Virtual Justice: The New Laws of Online Worlds, Yale University Press. Law, J., 2001. Notes on the Theory of the Actor Network'. Organizational Studies: Critical Perspectives on Business and Management, 5(4), p.853. Law, J. & Hassard, J., 1999. Actor Network Theory and After, WileyBlackwell. People's Republic of China, 2009. China bars use of virtual money for trading in real goods - joint statement by the Ministry of Culture and Ministry of Commerce of the People's Republic of China, Available at: http://english.mofcom.gov.cn/aarticle/newsrelease/commonnews/200906/20090606364208.html [Accessed March 21, 2011]. Reynolds, R., 2009. Competing Narratives in Virtual Worlds. In Third Person: Authoring and Exploring Vast Narratives. MIT Press, pp. 399-406. Reynolds, R., 2002. Intellectual Property Rights in Community Based Video Games. In The Transformation of Organisation in the Information Age: Social and Ethical Implications. University Lusada, pp. 455 - 470.

Reynolds, de Zwart

10

Stern v Sony Corporation of America, 2010. CV 09-7710 PA (FFMx), 8 February 2010., Taylor, T.L., 2002. Whose game is this anyway? Negotiating corporate ownership in a virtual world. In Computer Games and Digital Cultures Conference Proceedings. Citeseer, pp. 227-242. Williams, D., 2003. The Video Game Lightning Rod: Constructions of a New Media Technology, 19702000. Information, Communication and Society, 6(4), pp.523-550. de Zwart, M., 2010. Contractual Communities: effective governance of virtual worlds. University of New South Wales Law Journal, 33(2), pp.605 - 627. de Zwart, M., 2009. Piracy vs Control: Various Models of Virtual World Governance and their impact on Player Experience. Journal of Virtual Worlds Research, 2(3). Available at: https://journals.tdl.org/jvwr/article/viewArticle/663 [Accessed March 14, 2011].

Reynolds, de Zwart

11

The Game Behind the Video Game: Business, Regulation, and Society in the Gaming Industry New Brunswick, New Jersey, USA, April 8-9, 2011

Games are Not Coffee Mugs: Games and the Right of Publicity
William K. Ford The John Marshall Law School 315 S. Plymouth Court, Chicago Illinois, 60604 Raizel Liebler The John Marshall Law School 315 S. Plymouth Court, Chicago Illinois, 60604

ABSTRACT The right of publicity protects individuals from the unauthorized commercial exploitation of their identities. Should games be judged differently than other expressive mediums when it comes to this right? For approximately four decades, the dominant approach of the courts made games more vulnerable to right of publicity claims than more traditional mediums of expression, such as newspapers, magazines, books, films, and television programs. C.B.C. Distribution and Marketing, Inc. v. Major League Baseball Advanced Media, L.P., 505 F.3d 818 (8th Cir. 2007), upset the view that licensing [is] the settled order of things when it comes to games. (Petition for Writ of Certiorari, Major League Baseball Advanced Media v. C.B.C. Distribution and Marketing, Inc., (Feb. 22, 2008) (No. 07-1099), 2008 WL 515858 at 8). But uncertainty remains about how much flexibility game producers have to incorporate individuals into games without their consent. Commentators and courts are resisting a further weakening of the right of publicity beyond what C.B.C. accomplished. Some commentators think C.B.C. itself went too far. We argue that the value of free expression outweighs the value of licensing, meaning that the uncertainty in this area of the law should be settled by treating games in the same way as more traditional mediums of expression. KEYWORDS: Games, Law, Right of Publicity, First Amendment INTRODUCTION On October 16, 2007, the United States Court of Appeals for the Eighth Circuit upset the settled order of things when it comes to the right of publicity and games. (Petition for Writ of Certiorari, Major League Baseball Advanced Media v. C.B.C. Distribution and Marketing, Inc., (Feb. 22, 2008) (No. 071099), 2008 WL 515858 at 8). As a matter of state law, the majority rule in the United States is that using someones name or likeness (or identity) in advertising without that persons consent is a violation of the persons right of privacy or, using the more modern terminology, a persons right of publicity. Generally speaking, due to the interest of preventing conflicts with the First Amendment, no consent is required to use someones name or likeness for news or entertainment purposes in newspapers, magazines, books, television programs, or films. (ALI 1995: 47 cmt c). Some states laws potentially make fictional works infringing for incorporating the identity of a real individual, but courts usually resist finding infringement or find the First Amendment provides a defense. (ALI 1995: 47 cmt d; McCarthy 2011: 7:76; Cal. Civ. Code 3344(d); Winter v. DC Comics, 69 P.3d 473 (Cal. 2003)). Merchandising
Ford/Liebler 1

uses occupy a somewhat difficult middle ground between advertising uses and traditional expressive uses, but consent is usually required to use someones identity on a coffee mug or other item of merchandise like a poster or keychain. Where a literal celebrity likeness appears on a coffee mug or in a poster, the sum and substance of the expressive portion of the work is the celebrity likeness. The reason for someone buying the product, which is usually more costly than one without a celebrity likeness, is presumably the presence of the likeness rather than the mug or poster makers expressive contribution, for there is none. (Comedy III Prods v. Saderup, 21 P.3d 797, 809-810 (Cal. 2001)). What about games? Wittgenstein famously wondered what counts as a game, asking his readers to look and see whether there is something in common to all games, but concluding there is not. (Wittgenstein 1997: 66). Less famously, Bernard Suits disagreed: to play a game is to engage in activity directed towards bringing about a specific state of affairs, using only means permitted by rules, where the rules prohibit more efficient in favour of less efficient means, and where the rules are accepted just because they make possible such activity. (Suits 2005: 48-49). Most others seem to agree with Suits, at least about the rules. (Galloway 2006: 1). The games of interest to us are a subset of the games within Suits definition: commercial games sold in some physical or electronic form, such that one or more individuals name or likeness could feasibly be incorporated into it. Like newspapers, books, and films, many of these games communicate information and express ideas and emotion. They incorporate words, images, and/or music. But when words, images, and music are combined with a rule-governed process to create games, then what? Are they more like expressive works, such as books and films, or more like merchandise, such as coffee mugs and posters?1 As Ian Bogost (2007) emphasizes, the addition of processes need not undermine the expressive capacity of a work but can contribute to it. Nevertheless, until the Eighth Circuits decision in C.B.C. Distribution and Marketing and Marketing. v. Major League Baseball Advanced Media, 505 F.3d 818 (8th Cir. 2007), the safest prediction was that other than incidental uses (e.g., a trivia question in Trivial Pursuit,2 or uses of politicians names or likenesses in politically themed games), the courts would treat games like a coffee mug.3 Therefore, game publishers needed permission to use someones name or likeness in a game. This conclusion was reinforced by three cases from the late 1960s and early 1970s, each of which ruled against the game publisher defendants who used the plaintiffs identities without permission. However, in C.B.C., the Eighth Circuit found that the unlicensed use of Major League Baseball players names in fantasy baseball games violated Missouris right of publicity, but also that Missouri law conflicted with the First Amendment. (Id. at 823824). According to the defendant and numerous amici, the law was settled in favor of individuals controlling their appearances in games because games, like coffee mugs, are more commercially exploitive than informational or expressive. According to the settled view, the Eighth Circuit got it wrong. We disagree. The classic rule is wrong, or at least overbroad. DISCUSSION In the leading treatise on the topic, McCarthy describes the right of publicity as a right inherent to everyone to control the commercial use of identity and persona[.] (McCarthy 2011: 1:3). The Restatement 3d Unfair Competition defines it as the right to prevent the appropriation of the commercial value of a persons identity by using without consent the persons name, likeness, or other indicia of identity for purposes of trade[.] (ALI 1995: 46). According to McCarthys most recent count, 31 states
Some sources very casually link video games to merchandise. (Pollack (2011)). Karcher (2007: 570) argues Trivial Pursuit is less of a concern than fantasy baseball apparently because fantasy baseball is about present on-field performances and Trivial Pursuit is about the past. Its not clear to us why this distinction makes sense. It suggests trivia games should not use questions tied to closely to the present. Karcher also argues Trivial Pursuit takes less commercial advantage of individuals identities than fantasy baseball, which is another way of saying the use is incidental. 3 We are not aware of any litigation involving a politically-themed game, but given the courts reluctance to curb political speech and by analogy to politically themed posters, a right of publicity claim targeting a politicallythemed game would likely fail. (Paulsen v. Personality Posters, Inc., 59 Misc. 2d 444 (N.Y. Sup. Ct. 1968)).
2 1

Ford/Liebler

have recognized the right of publicity either by statute or common law or both. Only two states have rejected the right of publicity as a matter of common law, Nebraska and New York, but both of these states have statutory provisions. (McCarthy 2011: 6:3). Sometimes, as in New York, the right of publicity is labelled a right of privacy, a variation due to the fact that the right of publicitys origin lies in a broader right of privacy. (N.Y. Civ. R. 50; Warren & Brandeis 1890; Nimmer 1954; Prosser 1960). Indianas statute deserves special mention. For deceased individuals, courts generally apply the law of the deceased individuals citizenship at the time of death to determine whether the individual has a postmortem right of publicity. (McCarthy 2011: 11:8, 11:12, 11:15, 11:16). The right of publicity is a property right; therefore, whether an individuals right of publicity passes through his or her estate depends on whether the individuals state of citizenship recognizes a post-mortem right of publicity. New York, for example, does not recognize a post-mortem right of publicity. (N.Y. Civ. R. 50; Pirone v. MacMillan, Inc., 894 F.2d 579, 585 (2d Cir. 1990)). Indianas statute, like Washingtons,4 applies to acts occurring within the state regardless of a personalitys domicile, residence, or citizenship (Ind. Code 32-36-1-1-(a)) and can pass by operation of a states laws of intestate succession regardless of whether the state recognizes the property rights set forth in the statute (Ind. Code 32-36-1-16(6)).5 Moreover, the protection of the Indiana statute lasts 100 years, the longest defined period in the United States. (Ind. Code 32-36-1-8(a)). While federal law does not recognize a right of publicity, it does recognize an analogous unfair competition claim under the Lanham Act. (15 U.S.C. 1125(a)(1)). The critical difference between a right of publicity claim and a Lanham Act claim is that the Lanham Act requires a likelihood of consumer confusion. When the identity of an individual is used without permission in connection with some good or service, the Lanham Act requires a likelihood that consumers will be confused about whether the individual approved of the use. (15 U.S.C. 1125(a)(1)(A); Restatement 3d of Unfair Competition 46 cmt b). Disclaimers may protect against a successful Lanham Act claim (Food Scis. Corp. v. Nagler, No. 09-1798, 2010 U.S. Dist. LEXIS 112072, at *25-27 (D.N.J. Oct. 20, 2010)); they will not protect against a right of publicity claim. The starting point for the present discussion is what qualifies as a commercial use or a use for purposes of trade. Section 47 of the Restatement defines uses for purposes of trade as advertising uses or uses on merchandise. (ALI 1995: 47). The judicial opinion that coined the term right of publicity was about merchandise, i.e., baseball cards. (Haelan Laboratories, Inc. v. Topps Chewing Gum, Inc., 202 F.2d 866 (2d Cir. 1953)). In commenting on the merchandise category, the Restatement specifically mentions posters, buttons, and other memorabilia as examples of merchandising uses. (ALI 1995: 47 cmt b). The Reporters Notes to this section are more explicit, referring to games as examples of merchandising uses with citations to the three classic game cases, Palmer v. Schonhorn Enterprises, Inc., 96 N.J. Super. 72 (N.J. Super. Ct. 1967); Uhlaender v. Henricksen, 316 F. Supp. 1277 (D. Minn.1970);
A recent decision held that the Washington provision ignored a decedents state of domicile unconstitutional, violating Due Process Clause, the Full Faith and Credit Clause, and the Commerce Clause. (Experience Hendrix, L.L.C. v. Hendrixlicensing.com, Ltd., No. 09-285Z, 2011 U.S. Dist. LEXIS, at *41 n. 21, 42, 52 (W.D. Wash. Feb. 8, 2011)). The state of Washington was not a party to the litigation, however, and district court decisions are not binding in subsequent cases. (Starbuck v. San Francisco, 556 F.2d 450, 457 n.13 (9th Cir. 1977); Bilal v. Lehman, No. 04-2507, 2006 U.S. Dist. LEXIS 93152, at *16-17 (W.D. Wash. Oct. 2, 2006)). 5 Despite this language, decisions are split on the meaning. Without discussing the statutory language, one judge of the United States District Court for the Southern District of Indiana thought sixteen legendary Major League Baseball players state of domicile at death defines their rights of publicity. (CMG Worldwide, Inc. v. Upper Deck Co., No. 08-761, 2008 U.S. Dist. LEXIS 85497, *14 (S.D. Ind. Oct. 22, 2008) (Topps has not established that any of the Legends were domiciled in Indiana at the time of each of their respective deaths.). Another decision of the same district court explicitly held that the state of domicile at the time of death of a famous American cyclist, Marshall Major Taylor, was irrelevant under the statute. (Donovan v. Bishop, No. 09-275, 2010 U.S. Dist. LEXIS 110204, at *15-16 (S.D. Ind. Oct. 14, 2010) (Bishop claims that because Major Taylor died in Illinois and was apparently domiciled there at the time of his death, Illinois substantive law should apply. . . . The fact that Major Taylor died elsewhere is, according to the Statute, unimportant.).
Ford/Liebler 3
4

and Rosemont Enterprises, Inc. v. Urban Systems, Inc., 72 Misc.2d 788, 340 N.Y.S.2d 144 (Sup.Ct.), affd as modified 42 A.D.2d 544, 345 N.Y.S.2d 17 (1973). (ALI 1995: 47 Reporters Notes). While highly persuasive, the Restatement is usually not binding,6 and as a state law doctrine, the status of games could vary from state to state. No state statute, however, explicitly deals with games. Californias statutory right of publicity provisions refer to uses of someones identity on or in products, merchandise, or goods. (Cal. Civ. Code 3344(a), 3344.1(a)(1)). Illinois statute refers to an individuals right to control uses for commercial purposes (765 ILCS 1075/10), which is defined in part to include uses on or in connection with the offering for sale or sale of a product, merchandise, goods, or services ( 765 ILCS 1075/5). New Yorks statutes prohibit uses for purposes of trade without explicitly referencing merchandise. (N.Y. Civ. R. 50, 51). The lack of specificity in these and other state statutes leaves the question open for the courts as to whether games qualify as commercial uses or uses for purposes of trade. The three classic cases provided an initial answer decades ago, but only one of the three is a binding authority and then only in New York.7 Moreover, the expressive quality of the gaming medium is much more substantial than it was when these three cases were decided. Proponents of treating games as merchandise were perhaps too quick to declare the matter settled. Whether it makes sense to do so requires some consideration of the reasons for the right of publicity. The justification for the right of publicity is the subject of substantial controversy. The Restatement concedes that the arguments for protecting the right of publicity are weaker than the arguments for protecting other kinds of intellectual property. (ALI 1995: 46 cmt c). The dominant arguments are tied to either an economic analysis or a Lockean natural rights analysis. One economic argument is that the right of publicity creates incentives to create an identity, but this argument is not widely accepted. There are many benefits, including monetary benefits, to celebrity status separate from the revenue tied to the right of publicity. (Cardtoons, L.C. v. Major League Baseball Players Assn, 95 F.3d 959, 974 (10th Cir. 1996); Grady: 110-112 (1994)). Even if there is a small incentive effect, the social benefits tied to the cheaper exploitation of celebrity identities likely outweigh a small increase in the number of celebrities. In other words, more competitors offering fantasy baseball to consumers without passing on the cost of licensing fees is preferable to a tiny increase, in the fullness of time, in the number of professional athletes. An economic argument works much better for advertising. Separate from the value of endorsementsand false endorsements are already prohibited by the Lanham Actanother value of using a celebrity in an advertisement is to grab peoples attention. Overuse of a celebrity in advertising, a probable consequence if no consent is required, would drive the celebritys advertising value to zero. In terms of advertising value, the celebrity of the moment would quickly be used up. (Cardtoons, L.C. v. Major League Baseball Players Assn, 95 F.3d 959, 975 (10th Cir. 1996); Posner 1981: 258). Grady (1994: 119-120) extends this same reasoning to merchandising uses, worrying that overexposure to celebrities on t-shirts, posters, coffee mugs, and other similar uses will similarly dissipate the value of a
There are some exceptions, such as 1 V.I. Code 4 (2011): The rules of the common law, as expressed in the restatements of the law approved by the American Law Institute, and to the extent not so expressed, as generally understood and applied in the United States, shall be the rules of decision in the courts of the Virgin Islands in cases to which they apply, in the absence of local laws to the contrary. Whether a provision adopted in 1957 makes a restatement published in 1995 binding is an interesting question. (In re Kelvin Manbodh Asbestos Litig. Series, 47 V.I. 215, 227-237 (2005)). 7 Only the New York decision was affirmed by an appellate court. Decisions of federal district courts are not binding on other courts or even within the same district court. (Southeastern Stud & Components, Inc. v. Am. Eagle Design Build Studio, 588 F.3d 963, 967 (8th Cir. 2009); TMF Tool Co. v. Muller, 913 F.2d 1185, 1191 (7th Cir. 1990)). The same rule applies in New Jersey. (Raubar v. Raubar, 315 N.J. Super. 353, 358 n.3 (N.J. Sup. Ct. Ch. Div. 1998) (collecting cases); May Stores Shopping Centers, Inc. v. Hartz Mountain-Free Zone Center, 162 N.J. Super. 130, 136 (N.J. Sup. Ct. Ch. Div. 1978); Mazza v. Insurance Co. of North America, 149 N.J. Super. 60, 62 (N.J. Sup. Ct. Law. Div. 1977); Ferraro v. Ferro Trucking Co., 72 N.J. Super. 519 (N.J. Sup. Ct. 1962). The New York appellate decision in Rosemont Enterprises, however, is binding state-wide. (People v. Turner, 840 N.E.2d 123, 127 (N.Y. 2005); Mountain View Coach Lines, Inc. v. Storms, 102 A.D.2d 663, 664-665 (N.Y. App. Div. 1984).
Ford/Liebler 4
6

celebritys name or likeness. Grady offers no plausible examples of this occurring, and it is difficult to think of any examples where extensive celebrity merchandising began to threaten the value of the celebritys image. If Grady is right, then we should predict that in the rare cases where celebrities managed to saturate the market with merchandise, there was a threat of dissipation and rational actors managing the commercial uses of their identities acted to reduce the supply of merchandise. It is difficult, however, to identify even plausible examples. The famous 1976 Farrah Fawcett poster may be an example. (Bennetts 2009). Overexposure in tabloids seems more plausible, but courts will not restrict tabloid publishers from providing extensive or even excessive coverage of celebrities because the coverage is so clearly protected speech. Although many commentators rely on an economic analysis, a Lockean natural rights analysis is also clearly relevant in the recognition of the right of publicity. McCarthy (1987: 1711) suggests commentators are reluctant to rely on visceral feelings of fairness, but also suggests, courts should hesitate to ignore widely held beliefs even if it is difficult to find an economic rationale in support of them. Nimmer (1954: 216) put it this way: It would seem to be a first principle of Anglo-American jurisprudence, an axiom of the most fundamental nature, that every person is entitled to the fruit of his labours unless there are important countervailing public policy considerations. Many judicial decisions echo this view. (McFarland v. E & K Corp., No. 4-89-727, 1991 U.S. Dist. LEXIS 1496, at *4 (D. Minn. Jan. 17, 1991); Uhlaender, at 1282). As a related point, and one related to intellectual property generally, many courts state that an individual should not reap where they have not sown by, for example, free riding on the work of celebrities who have invested in their identities. (Lange & Powell 2009: 155; Intl News Serv. v. Associated Press, 248 U.S. 215, 239-240 (1918); State ex rel. Elvis Presley Meml Found. v. Crowell, 733 S.W.2d 89, 98 (Tenn. Ct. App. 1987)). Even if we accept that celebrities and other individuals are entitled to the fruit of their labours, as Grady (1994: 108-109) notes, this explanation for the right of publicity fails to explain cases where celebrities do not prevail even though someone has benefitted in some way from using their names or likenesses. This explanation may provide a starting point, but it does not assist with Nimmers qualification by explaining when there are countervailing policy considerations. On the other hand, the economic analysis poorly explains the application of the right of publicity rule for merchandise, suggesting both the economic and Lockean justifications contribute something to the analysis of the right of publicity. But other interests are at stake as well, including the protection of speech. Courts end up balancing these competing interests. Many games are expressive in ways that implicate First Amendment concerns. They have themes. They tell stories. Games often feature words, either written or spoken, which fits the literal scope of the text, but the courts interpret the First Amendment broadly to protect more than just words. (Tex. v. Johnson, 491 U.S. 397, 404 (1989)). Music, for example is protected, even without lyrics. (Ward v. Rock against Racism, 491 U.S. 781, 790 (1989); Reed v. Village of Shorewood, 704 F.2d 943, 950 (7th Cir. 1983). The Supreme Court has held that in determining whether conduct is sufficiently communicative to raise First Amendment concerns, the question is whether [a]n intent to convey a particularized message was present, and [whether] the likelihood was great that the message would be understood by those who viewed it. (Id. at 404, quoting Spence v. Wash., 418 U.S. 405, 410-411 (1974)). Early cases involving video games rejected First Amendment protection for them on the ground that they failed to communicate or express ideas (Marshfield Family Skateland, Inc. v. Marshfield, 450 N.E.2d 605, 609-610 (Mass. 1983)), but the recent spate of cases finding government restrictions on minors access to video games unconstitutional suggests video games have graduated to a recognized medium of expression, one equivalent to the more traditional forms of expression (Video Software Dealers Assn v. Schwarzenegger, 556 F.3d 950 (9th Cir. 2009); American Amusement Machine Assoc. v. Kendrick, 244 F.3d 572 (7th Cir. 2001)). But video games are not the only types of games that can be fairly described as sufficiently communicative or expressive to warrant First Amendment protection. This recommends caution in enforcing publicity rights against game publishers. Early American games illustrate a long tradition of expressive content in games. Game historians assume the earliest games available in the United States were imported from England. Early English games focused on education, such as Royal Genealogical Pastime (1791), a race across shield-shaped
Ford/Liebler 5

spaces for 52 monarchs. (Shea 1969: 60-61; Goodfellow 1998: 70). It is not surprising then that the topics of these early imports included travel, history, and geography. (Whitehill 1999: 118). The two earliest known games produced in the United States are Travelers Tour Through the United States (1822) and Travelerss Tour Through Europe (1822). The next known game published, The Mansion of Happiness (1843), illustrates the emphasis placed on moral education in early American games. Players attempted to make it to the Mansion of Happiness at the center of the board. Landing on a virtue sent the player forward; landing on a vice sent the player backward. The spaces labelled idleness, for example, sent players to poverty. Many other games used similar mechanics to teach the benefits of good behaviour and the penalties for bad behaviour. (Whitehill 1999: 119-122). The goal of Milton Bradleys first game, The Checkered Game of Life (1860), was to reach Happy Old Age through good behaviour rather than Ruin. (Shea 1969: 48). Many games published throughout the Nineteenth Century were educational, with history and geography being popular topics, but by the 1880s, current events became popular topics too. (Whitehill 1999: 125-127).8 With games incorporating trivia and current events, it is no surprise that game designers even in the Nineteenth Century used the names and likenesses of individuals. An early example is Bulls and Bears: The Great Wall Street Game (1883). Described as a subtle commentary on the making of financial empires at the publics expense, the game included caricatures of William Henry Vanderbilt, Jay Gould, and Cyrus Field, all three investors in the railroad industry. (Hofer 2003: 82). Autograph Authors (1886) was an educational card game incorporating contemporary authors, most prominently Mark Twain, who was depicted on the box. (Hofer 2003: 66). Admiral Winfield Scott Schley appeared in Schley at Santiago Bay (1899). (Hofer 2003: 100). Theodore Roosevelt appeared in Roosevelt at San Juan (1899). (Hofer 2003: 101). Whether any of these games licensed the use of these individuals names and likenesses is unclear, but it seems unlikely, especially for Bulls and Bears. At least some licensing did occur in the Nineteenth Century. Cooper (1995: 23-29) identifies three baseball games from the Nineteenth Century that were endorsed by players.9 Professional baseball player Charles Louis Zimmer (and possibly other players), for example, endorsed Zimmers Baseball Game (c. 1894). Parker Brothers begin licensing celebrity identities at least as early as the 1920s. In the 1920s, Parker Brothers licensed Charles Lindberghs name (or nickname) for a series of games. Around the same time, Parker Brothers also licensed the use of Eddie Cantors name and likeness for Eddie Cantors Tell It to the Judge (1936). (Orbanes 2004: 71-72; Whitehill 1999: 120). While it is likely that a significant norm of licensing names and likenesses for games predated the three classic games cases, these cases presumably solidified the legal argument in favour of treating games like other forms of merchandise and provided the basis for describing licensing as the settled order of things. The first case, Palmer v. Schonhorn Enterprises, Inc., 96 N.J. Super. 72 (1967), concerned a game apparently entitled Pro-Am Golf. We have not found an actual copy or images of this game outside this case and therefore know only what the opinion says about it. According to the opinion, the box of the game featured a drawing of a generic golfer and the following caption: PRO-AM GOLF GAME. 18 Championship holes. Profiles and playing charts of 23 famous golfers. Yardage ruler. Ball markers. Tee. Flag. Score cards. Dice. AS CHALLENGING AND EXCITING AS GOLF ITSELF. (Id. at 74). The names of the 23 golfers were not revealed on the box, but inside were 23 Profile and Playing Charts, each with biographical and factual information on the famous golfers. Among the 23 were Arnold Palmer, Gary Player, Doug Sanders, and Jack Nicklaus, the plaintiffs in the case. How the game is
As some games teach virtue, some will inevitably teach wickedness. The object of the notorious German game Juden Raus! (Jews Out!) is to be the first player to force six Jews from their stores and out of the city. (MorrisFriedman and Schdler 2003: 48). 9 Cooper (1995) identifies three Nineteenth Century and numerous Twentieth Century games as player endorsed, but it is unclear how he determined that the game publishers actually licensed the use of the players names and likenesses. In some cases, his discussion makes clear the use was licensed, but in other cases its unclear whether he just assumed the use of a players identity was licensed.
Ford/Liebler 6
8

played is unclear, including whether the golfers real-world statistics played a role in the game. The court held the defendant game company violated the golfers rights of privacy, rejecting the argument that the game should be able to make use of the golfers names, likenesses, and factual information in the same way a newspaper article or a biography could use their names without needing their consent. Why? According to the court, a person has the right to enjoy the fruits of his own industry. (Id. at 79). It is, said the court, unfair that one should be permitted to commercialize or exploit or capitalize upon anothers name, reputation or accomplishments merely because the owners accomplishments have been highly publicized. (Id.). Three years later, baseball board games suffered a similar fate. In Uhlaender v. Henricksen, 316 F. Supp. 1277 (D. Minn. 1970), the court considered two board games, Negamcos Major League Baseball and Big League Baseball. Each game used the names and statistics of 500 to 700 major league players, making some use of these statistics in the game. The defendants were offered a license but turned it down. The plaintiffs claimed misappropriation, which the court treated as a claim for infringement of the right of publicity. The court did not explain why games are different than news articles, but it did describe the game as a commercial use. As in Palmer, the Uhlaender court explained that celebrities invest considerable time to make their names valuable and are therefore entitled to the fruit of their labours. (Id. at 1282-83). After another three years, Rosemont Enterprises, Inc. v. Urban Systems, Inc., 72 Misc. 2d 788 (N.Y. Sup. Ct. 1973), completed the trilogy. This time it was not professional athletes, but Howard Hughes and his objection to the The Howard Hughes Game. The booklet included with the game said it is designed to teach players about the life and business ventures of Mr. Hughes. (The Howard Hughes Game (1972)). Toward that end, the booklet included not only the rules of play but also the background on various Hughes activities like his The Hughes Tool Company, the Gulf Brewing Company, and Hughes Aircraft. The booklet would surely be protected by the First Amendment, but this booklet came with a board, tokens, and other playing pieces. The result, according to the court, was that the defendants were not disseminating news or educating the public. Instead, the defendants were selling a commodity, a commercial product, an entertaining game of chance, the outcome of which is determined by maneuvering tokens on a game board by the throw of the dice. (Id. at 790). The court added that the use of Hughes name and biographical information was not legitimate to the public interest. The information was merely the medium used to market a commodity familiar to us all in its varied types and forms. (Id.) The court said it was not offering a hard and fast rule, adding that each case must be decided on its own merits, balancing the interest in free expression with the preservation of inviolate personality and property. But one of the factors to consider in the balance is the medium. Games were not a preferred medium. As already noted, these three decisions are endorsed by the Restatement. (ALI 1995: 47 Reporters Notes). While the current edition is silent on the question, earlier editions of McCarthys treatise also endorse these outcomes. McCarthy (2009: 7:30) noted, We are all familiar with childhood board games, puzzles, card games, and the like that are designed to teach as well as entertain. Nevertheless, he thinks the use of non-political celebrities should not be permitted because board games and wall posters featuring these celebrities are not traditional media in which ideas are conveyed and should usually be viewed as more exploitive than informational or educational. (Id.) McCarthy objects that people who criticize these cases mainly point to games about the political process. (Id.) Those that play games themselves are likely to be much more familiar with the variety of games available and would not characterize the medium primarily in terms of childhood board games, nor think of the few exceptions to juvenile games only in terms of political games. At least a subsection of gamers, video gamers, are demographically not children by 2006, adults constituted over sixty percent of the market. (Jenkins 2006: 203). It may also be relevant that the courts decided these three cases just before some of the most important developments in the history of gaming. The 1970s were a period of great advances in the gaming industry. While Charles S. Roberts published the first modern board wargame in 1954, it was in the 1970s that historical board wargaming experienced a period of explosive growth. These are not a small exception to a board game category
Ford/Liebler 7

dominated by familiar childrens games. While the sales of well-known board games like Monopoly, Clue, and Chutes and Ladders surely outnumber the sales of (probably close to all) individual historical wargame titles, the number of published titles is in the thousands, covering numerous eras and events, though World War II dominates the genre (Steambubble Graphics 2003). There are numerous computer wargames as well, though we lack any estimate of the number. These games allow players to recreate history in a stylized manner and numerous games use the identities of relevant military leaders. Pattons First Victory: Tunisia, 1943 (Decision Games 2010), a recently published example, includes a map of Tunisia and Eastern Algeria and includes pieces that represent the units that participated in the battle. Some playing pieces represent various military units. Others represent individual leaders, including United States General George S. Patton and German Field Marshall Erwin Rommel. Under Indianas statute, Pattons estate, which actively seeks licensing opportunities, potentially has a veto on the production of games like this through 2045. (Patton, Estate of, 2011).10 Even Rommel is apparently covered through 2044. While this particular game is fairly simple by the standards of the genre, all of these games sacrifice some degree of realism for playability. Indeed, according to one leading wargame designer, On a realism scale of 1 to 10, the highest possible rating we can hope for with paper and cardboard is a 2. (MacGowan 1980). Depending on the game, however, designers in the industry often go to great lengths to maximize the historical elements that can reasonably be incorporated into a game of this fashion and then debate the historical veracity of the results. In one exchange, for example, the designers and playtesters extensively debated whether one World War II game properly represented the moral of American units relative to other nationalities. (Collier 1983; Greenwood 1983; McNamara 1983; Mishon 1983). To varying degrees these games are one way among many to study historical conflicts. (Kirschenbaum 2009: 360). Recognizing this, the National World War II Museum now incorporates an annual wargaming event into their schedule. (Natl WWII Museum 2011). The 1970s also saw the invention of a new genre of game, the tabletop role-playing game. Gary Gygax and Dave Arneson released Dungeons & Dragons in January 1974, creating the commercial roleplaying game industry. (Gygax 1987: 20). Cover (2010: 168) defines these games as a type of game/game system that involves collaboration between a small group of players and a gamemaster through face-to-face social activity with the purpose of creating a narrative experience. One book claims to catalogue over 250 different tabletop role-playing game systems and their even more numerous supplements. (Schick 1991). Dungeons & Dragons uses a fantasy setting, thereby reducing the relevance of incorporating living individuals into the game, but numerous role-playing game systems use settings where actual individuals would be relevant. As one example, GangBusters (TSR 1982), a 1920s roleplaying game, provides statistics for assorted criminals and law enforcement agents like Al Capone, John Dillinger, and Elliot Ness. Dillinger, who is used in several other games, including Dillinger (BSO Games 2002), The Godfather (E.A. 2006), The Godfather II (E.A. 2009), and Amazing Heists: Dillinger (Gamers Digital 2009), is notable because he is not only covered by the Indiana statute through 2034, but his estate also actively seeks to enforce his post-mortem right of publicity. (Gamers Digital, Inc. v. Dillinger, LLC, No. 09-3204 (D. Minn. Nov. 17, 2009) (complaint); Electronic Arts, Inc. v. Dillinger, LLC, No. 09-3965 (N.D. Cal. Aug. 27, 2009) (complaint); Phillips v. Scalf, 778 N.E.2d 480 (Ind. Ct. App. 2002). As one last major development, the 1970s were the beginning of the video game industry. Ataris Pong (1973) was not the first video game, but it was the first video game blockbuster. It demonstrated the commercial appeal of video games, paving the way for thousands of games. (Donovan 2010: 29-37
Pattons family, specifically his grandchild, is registered as the claimant for General Pattons publicity rights with the California Secretary of State. (California Registration of Claim as Successor-in-Interest, File No. 2002-001 for General George S. Jr. Patton (Jan. 25, 2002)). California provides a post-mortem right of publicity for 70 years after an individuals death. (Cal. Civ. Code 3344.1(g)). The Patton family history published by Robert H. Patton strongly implies the producers of George C. Scotts Patton (20th Century Fox 1970) did not obtain permission from the estate, but that had the producers done so, the estate would have refused. (Patton 1994: 295296). Prevailing against the producers of a film as opposed to a game, however, would be unlikely. (Cal. Civ. Code 3344.1(a)(2)).
Ford/Liebler 8
10

Campbell-Kelly 2003: 269, 272). Early video games were very simple in design, which as already noted led judges to doubt their expressive qualities in the 1980s, but video games eventually went well beyond Pong and other early games in terms of expressive qualities. Some new releases remain simple, even today, but numerous games provide elaborate stories and narrative structures. An entire book of essays is devoted to the meaning of Grand Theft Auto. (Garrelts 2006). Grand Theft Auto is an exceptional example, but many mundane video game feature significant expressive elements, as numerous courts have noted in responding to constitutional challenges to various state restrictions on violent video games. Comparing games generally to items like coffee mugs and posters seems somewhat absurd. Games offer vastly different options for interacting with ideas and information. Perhaps if the courts of the late 1960s and early 1970s better understood the expressive capacity of games, the cases would have come out differently. An opportunity to reconsider the question finally came in 2005. In the C.B.C. case, C.B.C. Distribution and Marketing, Inc. sought a declaratory judgment from the United States Court of Appeals for the Eighth Circuit that it could use Major League Baseball players names and statistics without a license in operating fantasy baseball games. The parties agreed Missouri law applied, and the court did not question this point. The court agreed C.B.C. infringed the players rights of publicity under Missouris right of publicity statute, but it further held the First Amendment provided a defense. In a relatively brief discussion that did not refer to the three classic game cases, one of which is from a district court in the Eighth Circuit, the court said C.B.C. engaged in protected expression by providing an interactive form for the use of the players information. Fantasy baseball is very closely tied to real world events and involves grinding through statistics, which probably elevates its status as something involving the dissemination of news and information. The court added that baseball players are highly paid, a fact that seems odd to rely upon because the right of publicity is usually of primary interest to celebrities. The court also noted that consumers are unlikely to be confused about the players endorsing C.B.C., which isnt even a requirement for infringement. (C.B.C. at 820, 823-824). After the C.B.C. decision, a similar result followed in a case involving the National Football League players. (CBS Interactive, Inc. v. NFL Players Assn, 259 F.R.D. 398, 419 (D. Minn. 2009)). The C.B.C. opinion is underwhelming, but the court was right to focus on the interactive use of information. One plausible response to balancing the right of publicity against the First Amendment is the transformative use test adopted in California, provided it is properly applied. In two cases, the California Supreme Court dealt with two ends of the spectrum. In Comedy III Productions v. Saderup, 21 P.3d 797 (Cal. 2001), the work was a conventional charcoal drawing of The Three Stooges. In Winter v. DC Comics, 69 P.3d 473 (Cal. 2003), the work was a Jonah Hex comic book with two half-human, half-worm characters based on two musicians, Johnny and Edgar Winter. According to the California Supreme Court, When artistic expression takes the form of a literal depiction or imitation of a celebrity for commercial gain, directly trespassing on the right of publicity without adding significant expression beyond that trespass, the state law interest in protecting the fruits of artistic labor outweighs the expressive interests of the imitative artist. . . . On the other hand, when a work contains significant transformative elements, it is not only especially worthy of First Amendment protection, but it is also less likely to interfere with the economic interest protected by the right of publicity. (Comedy III at 808 (emphasis added)). The courts reference to adding expression as a way to satisfy the transformative use test is critical. Elsewhere the court explained, Another way of stating the inquiry is whether the celebrity likeness is one of the raw materials from which an original work is synthesized, or whether the depiction or imitation of the celebrity is the very sum and substance of the work in question. We ask, in other words, whether a product containing a celebritys likeness is so transformed that it has become primarily the defendants own expression rather than the celebritys likeness. (Id. at 809). The court further clarified that the reference to expression means additional expression of something other than the likeness of the celebrity. (Id.). A poster with a conventional celebrity image lacks any expression beyond the image itself. The image is neither transformed itself nor part of a larger work that makes the celebrity image raw
Ford/Liebler 9

material for a larger work. Based on this test, the court ruled the First Amendment did not protect Saderups portrayal of the Three Stooges but it did protect DC Comics portrayal of the Winter brothers. The sum and substance of Saderups drawing was The Three Stooges. There was nothing more. The sum and substance of the comic book, however, went far beyond just the Winter brothers. The brothers were altered and surrounded by a larger story.11 Recently, courts have pushed back against the loosening of the right of publicity started in C.B.C., and they have done so in reliance on the California Supreme Courts transformative use test. In No Doubt v. Activision Publishing, Inc., 192 Cal. App. 4th 1018 (2011), the court held that Activisions use of No Doubt in Band Hero (2009) did not pass the transformative use test because the creative elements were only literal, fungible reproductions of [the bands] likenesses. (Id. at 1018). Although the court acknowledged that the context in which a celebritys likeness is supposed to be relevant to the transformative use test, it nevertheless zeroed in specifically on the depiction of No Doubt: In Band Hero . . . no matter what else occurs in the game during the depiction of the No Doubt avatars, the avatars perform rock songs, the same activity by which the band achieved and maintains its fame. Moreover, the avatars perform those songs as literal recreations of the band members. That the avatars can be manipulated to perform at fanciful venues including outer space or to sing songs the real band would object to singing, or that the avatars appear in the context of a video game that contains many other creative elements, does not transform the avatars into anything other than exact depictions of No Doubts members doing exactly what they do as celebrities. (Id. at 1034). Presumably, a documentary film maker would not need to turn the members of No Doubt into half-worm, half-human characters to report on the band. The reason a game is treated in this way is likely found in another section of the opinion where the court contrasted a video game with an artistic work, something that is conventional speech. (Id. at 1040). The critical issue of how the transformative use test will be applied in games is now being further tested in litigation in the federal courts involving amateur athletes and Electronic Arts NCAA football games. (Keller v. Elec. Arts, Inc., No. 09-1967, 2010 U.S. Dist. LEXIS 10719 (N.D. Cal. Feb. 8, 2010); In re Student-Athlete Name & Likeness Litig., No. 2212, 2011 U.S. Dist. LEXIS 11394 (J.P.M.L. Feb. 4, 2011); Hart v. Elec. Arts, 740 F. Supp. 2d 658 (D.N.J. 2010)). The plaintiff in Keller is a former college football player who seeks class action status to challenge EAs use of players in the various NCAA titles who match the descriptions of their real-world counterparts in terms of jersey numbers, height, weight, and state of origin. While the players are not given their real-world names, consumers can name the players themselves with rosters widely available for download on third-party websites. (Id. at *6-7). The district court held the uses of the players identities was not sufficiently transformative as a matter of law to support a motion to dismiss. The court said EA does not depict the plaintiff in a different form; he is represented as what he was: the starting quarterback for Arizona State University. (Id. at *16). Moreover, the setting is a football field. While EA urged the court to look at the work as a whole, the court claimed the Winter decision (and Kirby) show that this Courts focus must be on the depiction of Plaintiff in NCAA Football, not the games other elements. (Id. at *18.). To the extent the court truly believes context does not count, the consequences for uses of a celebritys name are striking. Surely a fictional entertainment broadcast that included explicit references to Samuel Michael Keller would be protected by the First Amendment. Despite a literal use of Kellers name rather than a transformed (and perhaps barely recognizable) version, a court would presumably look at the larger context and recognize Keller is not the sum and substance of such a work. In the wide range of uses of celebrity identities from Comedy III to Winter, at least games with celebrity images must be very close to Winter to qualify
Given this approach, an intermediate California appellate court was right to conclude the First Amendment protected Sega of America from a right of publicity claim brought by Kierin Kirby a.k.a. Lady Miss Kier. (Kirby v. Sega of America, Inc., 144 Cal. App. 4th 47 (2006)). Although there were arguably traces of Lady Miss Kier in Segas Space Channel 5, there were significant differences, including the 25th Century setting. (Id. at 59-60). Kirby is near the Winter end of the spectrum and Sega prevailed.
Ford/Liebler 10
11

for a First Amendment defense. Applied consistently to uses of identity, Patton cannot be literally depicted in a historical wargame, by name or image, despite the transformative context of the game as a whole. Nor can Al Capone or John Dillinger be literally depicted in a historical role-playing game. Their identities must be transformed, thereby undermining the medium of many opportunities to inject realism into games involving recent history or the present. CONCLUSION The public has come to accept an unlicensed use of a celebritys identity in a newspaper, magazine, book, television program, or movie, even if highly critical or unflattering, as quite normal and routine. Add a rule-governed process to the material contained in these other mediums, however, and the result is viewed as improper commercial exploitation. According to Karcher (2007: 571), nobody would suggest that players should not have the right to be compensated for the use of their identities in the video game . . . markets[.] We are among the nobodies. Comparing games to coffee mugs and other celebrity trinkets makes little sense. Games are often highly expressive and should be treated as such rather than as a second class medium of communication. When states attempt to extend the right of publicity to make game uses infringing, the First Amendment should usually prevent them from doing so. REFERENCES American Law Institute. (1995) Restatement (Third) Unfair Competition. Barton, Matt. (2008). Dungeons & Desktops: The History of Computer Role-Playing Games. AK Peters. Bennetts, Leslie with Becker, Jonathan. Beautiful People, Ugly Choices. Vanity Fair, Sept. 2009, 302. Bogost, Ian. (2007). Persuasive Games. MIT. Campbell-Kelly, Martin. (2003). From Airline Reservations to Sonic the Hedgehog: A History of the Software Industry. MIT. Collier, J. M. (1983). Glass Anvil: A Dissenting View of G.I. Anvil of Victory. The General, May-June, 24-27. Cooper, Mark. (1995). Baseball Games. Schiffer. Cover, Jennifer Grouling. (2010). The Creation of Narrative in Tabletop Role-Playing Games. McFarland. Donovan, Tristan. (2010). Replay: The History of Video Games. Yellow Ant. Galloway, Alexander R. (2006). Gaming: Essays on Algorithmic Culture. Univ. Minnesota. Garrelts, Nate. (2006). The Meaning and Culture of Grand Theft Auto. McFarland. Greenwood, Don. (1983). The G.I. Design Team Replies. The General, May-June, 28-29. Grady, Mark F. (1994). A Positive Theory of the Right of Publicity. 1 UCLA Entertainment Law Journal 97-126. Goodfellow, Caroline G. (1998). The Development of the English Board Game, 1770-1850. 1 Board Game Studies, 70-80. Gygax, Gary. (1987). Role-Playing Mastery. Perigee. Hofer, Margaret K. (2003). The Games We Played. Princeton Architectural Press. Karcher, Richard T. (2007). The Use of Players Identities in Fantasy Sports Leagues: Developing Workable Standards for Right of Publicity Claims. 111 Penn. St. Law. Review 557-585. Kirschenbaum, Matthew. (2009). War Stories: Board Wargames and (Vast) Procedural Narratives. In Pat Harrigan and Noah Wardrip-Fruin (Eds.) Third Person: Authoring and Exploring Vast Narratives Jenkins, Henry. (2006). Fans, Bloggers, and Gamers: Exploring Participatory Culture. New York Univ. Press. Lange, David L. and Powell, H. Jefferson. (2009) No Law: Intellectual Property in the Image of an Absolute First Amendment. Stanford. MacGowan, Roger. (Sept.-Oct. 1980). F&M Interview John Hill All American. Fire & Movement, 42, 44.

Ford/Liebler

11

McCarthy, J. Thomas. (1987). Melville B. Nimmer and the Right of Publicity: A Tribute. 34 UCLA Law Review 1703-1712. McCarthy, J. Thomas. (2009). The Rights of Publicity and Privacy. West. McCarthy, J. Thomas. (2011). The Rights of Publicity and Privacy. West. McNamara, Bob. (1983). Rechecking Our Sources. The General, May-June, 29-33. Mishcon, Jon. (1983). Speaking from the Playability Viewpoint. The General, May-June, 29. Morris-Friedman, Andrew and Schdler, Ulrich. (2003). Juden Raus! (Jews Out!)Historys Most Infamous Board Game. 6 Board Game Studies 47-57. Natl WWII Museum. (2011). Heat of Battle V Wargaming Convention, available at http://www.nationalww2museum.org/calendar/educational-wargaming-at-the.html. Nimmer, Melville B. (1954). The Right of Publicity. 19 Law & Contemporary Problems 203-223. Orbanes, Philip E. (2004). The Game Makers. Harvard Business School. Patton, Estate of General George S. Jr., Business and Licensing Opportunities, available at http://www.generalpatton.com/business.html. Patton, Robert H. The Pattons: A Personal History of an American Family. Brasseys. Pollack, Malla. (2011). Litigating the Right of Publicity: Your Clients Face Was on the News, Now Its on T-Shirts and Video Games, 119 American Jurisprudence Trials 343. Posner, Richard. (1981). The Economics of Justice. Harvard. Prosser, William L. (1960). Privacy. 48 California Law Review 383-423. Schick, Lawrence. Heroic Worlds: A History and Guide to Role-Playing Games. Prometheus. Shea, James J. (1960). Its All in the Game. G.P. Putnams Sons. Steambubble Graphics (2003). The Games Database, available at http://www.isimulacrum.com. Suits, Bernard. (2005). The Grasshopper: Games, Life and Utopia. Broadview. Warren, Samuel D. and Brandeis, Louis D. (1890). The Right to Privacy. 4 Harvard Law Review 193220. Whitehill, Bruce. (1999). American Games: A Historical Perspective. 2 Board Games Studies 116-141. Wittgenstein, Ludwig. (1997). Philosophical Investigations. (trans. G.E.M. Anscombe). Blackwell.

Ford/Liebler

12

The Game Behind the Video Game: Business, Regulation, and Society in the Gaming Industry New Brunswick, New Jersey, USA, April 8-9, 2011

New Media Censorship & the First Amendment: Understanding the Origins of EMA v. Schwarzenegger
Thomas H. Rousse Northwestern University 1104 Garnett Pl, Evanston, IL tommy.rousse@gmail.com

ABSTRACT Schwarzenegger v. EMA is not the first time the regulation of electronic games has come before the judicial system. This paper explores the early case history of electronic games and questions the origins and usefulness of the informational standard that became the basis for widespread electronic game regulation in the 1980s. By examining contemporary games and the relevant court cases, both the informational theory and its application are critiqued. This instructive example of how the American legal system deals with the First Amendment status of new media points to a need for a more robust definition of speech. KEYWORDS: Schwarzenegger v. EMA, First Amendment, new media, electronic games, regulation, Missile Command

INTRODUCTION:
Schwarzenegger v. EMA is poised to be the most significant case involving the free speech categorization of electronic games. While a Supreme Court ruling in favor of electronic games as fully protected speech would be a major victory for the medium and its supporters, it may conceal a larger problem: novel media which are eventually designated as protected speech often pass through a significant period of censorship or uncertainty about their legal status as expression. In the case of film, from the earliest commercial exhibitions to the Burstyn v. Wilson decision in 1952, that period was almost sixty years, while electronic games have gone a little over three decades without a definitive ruling. If a medium is deemed worthy of protection and no significant changes in its structural capabilities have occurred (certainly, film did not undergo a technical revolution in 1952 that suddenly made it expression), then a significant period exists where rightfully protected speech can and will be restricted or censored. This problem can be significantly exacerbated by the fledgling state of many new media industries: their lack of trade organizations or legal resources can lead to weak legal cases, which in turn establish precedent for ruling against the expressive capacity of new media. Electronic games, from their case history beginning in the early 1980s, exhibit exactly this pattern. Municipal and state courts, confronting local ordinances contested by small business owners in the arcade business, establish a set of precedents with little coherence and scant precedence in a newfound informational standard for what constitutes speech. This paper will recount the early case history of electronic games ultimately leading up to the Schwarzenegger v. EMA decision and will carefully analyze the informational standard and its deficiencies.

Rousse

While video games attracted negative attention from media outlets and parents shortly after they became commercially available, it would take a few more years for the emerging medium to become the subject of legal regulation. Censorship and regulation often take place without government intervention. Nonetheless, government regulation and censorship can be particularly pernicious due to its potential effectiveness and legitimacy. Through the majesty of its authority and the punitive threat of monetary fines or imprisonment, the government from the municipal to the federal level can stifle expression in ways that few other actors can. At stake is more than the freedom to play video games: a Supreme Court ruling that denies electronic games First Amendment protection could have far-reaching consequences. Future media may be stunted in their development or stifled altogether if such a ruling is taken as precedent that traditional media warrants more protection than novel forms. Regulation of violent content in video games may lead to similar regulation in film, television, and print. The legal battle for the status of video games does not take place in a vacuum; ultimately, the whole of First Amendment tradition can be sullied or reaffirmed by the legal reasoning behind electronic games-related cases. By 1982, as coin-operated arcade machines became ubiquitous, cities and towns passed ordinances to restrict the use of early electronic games. These regulations ranged from restrictions on the location of arcades and their hours of operation to outright bans on all coin-operated machines, including video games. Arcade operators and others contested these regulations by claiming that electronic games should be protected as speech under the First Amendment. This point of contention would remain at the heart of litigation on electronic games for the next two decades. Since expression within the ambit of the First Amendment is considerably more difficult to regulate than conduct or expression outside of it, the categorization of electronic games as speech is one of the fundamental considerations for determining permissible regulations or censorship. When electronic games are recognized as speech, only a few regulations meeting stringent guidelines will pass constitutional muster. When they are not, legislators must meet few requirements when drafting regulations or bans. As the complexity, subject matter, and expressive capacity of the medium has developed and public attitudes have changed, the classification of electronic games has shifted and still remains controversial. The majority of courts have gradually recognized electronic games as speech after declaring them ineligible for such protection during the mediums formative years. But not all legal scholars are convinced they have made the right decision, and the Supreme Court has yet to weigh in on the matter. Without a clear definition of speech, rightfully protected expression in the form of electronic games was vulnerable to restriction for nearly three decades. This lack of a useful definition has led to a shaky enough precedent for the appropriate protection of electronic games that cities and states around the U.S. still periodically pass laws which either fail to meet constitutional standards or are currently on appeal to a higher court.

Mereentertainment?Earlyregulationofelectronicgames
The issue of whether or not electronic games are constitutionally protected speech first received court attention in Americas Best Family Showplace Corporation v. City of New York Department of Buildings, heard by the Federal District Court for the Eastern District of New York in 1982.1 Americas Best Family Showplace planned to open a restaurant which would feature video game screens at each table. A previous ordinance restricted the number of arcade machines in a business location to four unless the operators received a license for an arcade, a process that was essentially impossible to complete.2 Americas Best Family Showplace sought a ruling that the restriction violated the First and Fourteenth Amendment rights of the arcade operator and his customers.3

536 F. Supp. 170 (E.D.N.Y. 1982) Hereafter, Americas Best. Ibid, 172. 3 Ibid, 171-172. N.B. Through the Fourteenth Amendment, states and local governments must also abide by the First Amendment.
2

Rousse

The court recognized that the case hinged on whether or not video games receive First Amendment protection as speech: A dispositive threshold question is whether video games are speech or expression protected by the First Amendment. If they are not, the likelihood of plaintiffs success on the merits at trial is minimal, given the traditionally heavy presumption of constitutionality accorded to a local governments exercise of its zoning powers.4 In footnote four of United States v. Carolene Products (1938), the Supreme Court established that most city ordinances are presumed constitutional if they have a rational basis: if a legislature presented evidence supporting a relationship between the regulation and its intended goal, or even if it had not but such a relationship could be reasonably assumed by the court, the regulation would be upheld. If that regulation dealt with a fundamental right such as those protected by the First Amendment, however, it would have to meet the strict scrutiny standard.5 Under strict scrutiny, the government must overcome a prohibitively high standard of evidence, intended to protect expression and other civil liberties from all but the most carefully considered and empirically-supported regulations. When dealing with a right such as speech, government regulation or censorship of speech is assumed to be invalid. To overcome this presumption, the government must provide extensive evidence connecting the means of regulation and its intended goal. The courts independently examine this evidence to determine if the regulation is narrowly tailored and serves a compelling state interest.6 For a regulation to be valid under strict scrutiny, several conditions must be satisfied: first, it must restrict protected speech as little as possible; second, the government must demonstrate that this specific regulation is necessary to advance a compelling state interest; third, no less restrictive alternatives can exist. This threshold is so difficult to reach that strict scrutiny in free speech, particularly when applied to suspect classes of vulnerable groups like racial minorities, has been described as strict in theory, but fatal in fact.7 The plaintiffs case in Americas Best relied on Schad v. Borough of Mt. Ephraim (1981)8 to establish electronic games as protected speech under the same criteria accepted by the Schad court. In Schad, the Supreme Court decided that a usually nude dancer behind a coin-operated view hole was constitutionally protected speech. Under the strict scrutiny standard, the evidence that Mt. Ephraim provided in Schad was found to be insufficient to show that the ban on live entertainment was narrowly tailored enough to pass constitutional muster. Notably, the Americas Best court also reiterates a point made in Schad: Dancing, in all its non-obscene forms, has long been considered a form of expression protected by the First Amendment.9 This emphasis on a form of expression supports the media-centric approach of the expressive germ perspective. This is distinct from a method that would require each individual example of a format to demonstrate some prerequisite to qualify for protection. The Americas Best court then invokes Joseph Burstyn, Inc. v Wilson (1952),10 the landmark case which established film as protected speech, as a standard to determine the speech status of electronic games. They quote:
It cannot be doubted that motion pictures are a significant medium for the communication of ideas. They may affect public attitudes and behavior in a variety of ways, ranging from direct espousal of a political or social doctrine to the subtle shaping of thought which characterizes all artistic expression. The

Ibid, 173. United States v. Carolene Products, 304 U.S. 144, (1938). 6 R.A.V. v. City of St. Paul, 505 U.S. 377, 395. (1992) 7 But cf. Winkler, Adam. Fatal in Theory and Strict in Fact: An Empirical Analysis of Strict Scrutiny in the Federal Courts. Vanderbilt Law Review, Vol. 59, p. 793, 2006; Winkler finds that laws survive Supreme Court strict scrutiny 30% of the time on average (but local ordinances survive only 17% of cases). 8 452 U.S. 61. 9 Americas Best, footnote 11. 10 343 U.S. 495, 72 S. Ct. 777.
5

Rousse

importance of motion pictures as an organ of public opinion is not lessened by the fact that they are designed to entertain as well as to inform.11

The court includes in this quotation the Burstyn courts quote of Winters v. New York (1948):12 The line between the informing and the entertaining is too elusive for the protection of that basic right [a free press].13 They neglect to quote the sentence in Winters directly proceeding the Burstyn quotation, which states: We do not accede to appellees suggestion that the constitutional protection applies only to the exposition of ideas.14 Nonetheless, the Americas Best court asserts: it seems clear that before entertainment is accorded some First Amendment protection there must be some element of information or some idea being communicated.15 On this standard, the court concludes that video games are not protected speech: In no sense can it be said that video games are meant to inform. Rather, a video game, like a pinball game, a game of chess, or a game of baseball, is pure entertainment with no informational element.16 Since video games were not found by the court to qualify for First Amendment protection, the ordinance and its justification met the reduced requirements of the rational basis test for constitutional regulation.17 The genesis of the standard adopted by the Americas Best court is unclear at best. Neither of the cases cited in the decision - Joseph Burstyn, Inc. v. Wilson (1952) and Southeastern Promotions, Ltd. v. Conrad (1975)18 - establish the capacity to transmit ideas or information as the sole definition or test of protected speech. The word information does not even appear in Southeastern Promotions; idea is present only in a dissent, and has no relationship to the test adopted in Americas Best. The segments of Southeastern Promotions cited deal mainly with prior restraint of speech, not determining whether or not the material at issue (the Broadway musical Hair) contains information or ideas. Burstyn, while clearly establishing film as a protected format of speech and citing it as a significant medium for the communication of ideas, does not use information as the sole determination of a formats status as protected speech.19 The Americas Best reading of the Burstyn standard fixates on information and jettisons the rest of the quotation they cite for support. Beyond information, the Court recognizes in Burstyn that films may affect public attitudes and behavior from direct espousal of a political or social doctrine to the subtle shaping of thought which characterizes all artistic expression.20 A careful reading of the final sentence of the Burstyn quotation does not exclude pure entertainment or establish information as the sole arbiter of protected status: The importance of motion pictures as an organ of public opinion is not lessened by the fact that they are designed to entertain as well as to inform. 21 It simply says that mixing entertainment and information does not remove constitutional protection, not that information is the only important element of protected speech, which may be adulterated but not discounted by the presence of entertainment. The only justification for the informational capacity test given in the Americas Best decision is that it seems clear to the court that this is a component of protected speech. This standard is no more sustainable than Justice Potter Stewarts infamous I know it when I see it, 22 definition of obscenity.

11 12

Americas Best, 173-174, quoting Joseph Burstyn, Inc. v. Wilson, 343 U.S. 495, 72. (1952) 333 U.S. 507 (Hereafter, Winters.). 13 Americas Best 173. 14 Winters, 65. 15 Americas Best, 173. 16 Ibid, 174. 17 Americas Best, 176. 18 420 U.S. 546. Hereafter Southeastern Promotions. 19 Burstyn, 501. 20 Ibid. 21 Americas Best, 173-174, quoting Joseph Burstyn, Inc. v. Wilson, 343 U.S. 495, 72. (1952) 22 Jacobellis v. Ohio. 378 U.S. 184, Stewarts concurrence. Rousse 4

Video games received similar treatment in Caswell v. Licensing Commission for Brockton (Mass. S.J.C. 1983).23 As a local business owner, Caswell wanted to open an arcade featuring seventy machines. The city planning commission denied his proposal, citing concerns about pedestrian safety and juvenile loitering. Caswell claimed that this restriction on electronic games violated the First Amendment because of the mediums similarity to television and film, stressing in particular that electronic games are capable of articulating a plot. To support this assertion, Caswell submitted Space Invaders to the record. As seen in Americas Best, the Caswell court used the standard of the communication of ideas for First Amendment protection, emphasizing the physical element of video game interaction and comparing it to recreational rollerblading. The court found that Space Invaders, Caswells example of the expressive power of a video game, did not meet the information requirement established by Americas Best.24 Thus, the denial of his proposal was found constitutional. Because the court found that no fundamental right was involved, Caswells contention that the ordinance was unconstitutionally vague and violated his freedom of assembly and association was also denied.25 The connection between the cases the Caswell court cites and the informational standard formerly seen in Americas Best is not apparent. In fact, some of the specific page citations contain segments that seem to directly contradict the courts narrow construction of information as the sole prerequisite for First Amendment protection, such as this excerpt from page 65 of the Supreme Courts decision in Schad v. Mt Ephraim (1981): Entertainment, as well as political and ideological speech, is protected .26 A careful review of their respective texts reveals that none of the other cited cases27 clearly establish the information test used in the Americas Best or Caswell decisions; many mention ideas or information obliquely at best, and no form of speech is given or denied constitutional protection based on such a test. While the previous two cases concern the denial of license to individual businesses, City of Warren v. Walker (Mich. App. 1984) focused on a citywide regulation that restricted individuals under seventeen from playing a video game in public unless a parent or guardian was present.28 The City of Warren court searched in vain to find a justification or precedent that qualified electronic games as protected speech, but ruled that they did not possess the expressive element established in other cases.29 As indicated by the oddly wistful description of their own deliberation (i.e. searched in vain), this court, while coming to the same conclusion as Caswell and Americas Best, was more hesitant to rule that video games lacked any expressive elements. In a lengthy footnote, the court recognizes a tradition of youth recreational activities restricted for their purported immoral influence represented by the Trouble in River City scenario from The Music Man, but this observation does not reach the constitutional question.30 Furthermore, the court acknowledged that video games may already possess some expressive capacity and may develop the requisite informative capacity to be considered speech. The plaintiff, however, did not present any video game to the court that served as an example of this capacity but argued instead that the format was already an accepted category of protected expression. The City of Warren Court found no precedent to support that claim.31 These cases provided a solid precedent for courts to deny video games First Amendment protection for the rest of the decade with little pause. There are no significant cases that reached a district court or further that considered a different conclusion for electronic game until 1991s Rothner v. City of
23 Caswell v. Licensing Commission for Brockton, 444 NE 2d 922 Mass: Supreme Judicial Court, Plymouth, 1983. Hereafter, Caswell. 24 Caswell, 869-871. 25 Ibid, 872-74. 26 452 US 61 27 Doran v. Salem Inn, Inc. 422 U.S. 922 (1975), Erznoznik v. Jacksonville, 422 U.S. 205 (1975), Southeastern Promotions, Ltd. v. Conrad, 420 U.S. 546 (1975), and Schacht v. United States, 398 U.S. 58. 28 City of Warren v. Walker. Court of Appeals of Michigan, 135 Mich. App. 267, 1984. Pg. 270. Hereafter City of Warren. 29 Ibid, 273. 30 Ibid, 274, footnote 2. 31 Ibid, 275, footnote 3.

Rousse

Chicago.32 Cities and towns were required to meet only the rational basis test for the regulation of video games because they found no evidence that video games met the requirements of protected expression. Questions remain as to whether the standard to which electronic games were held was appropriate or if the evidence presented did justice to the state of the medium.

Areelectronicgamesspeech?
While the cases reviewed here are unanimous in their rejection of electronic games as protected speech, not all of the courts were confident in this characterization. We should consider, first of all, that the weighty issue of establishing an entire format representing hundreds of games and billions of dollars as constitutionally protected was left to individual business operators to litigate. No industry-wide establishment charged with representing the legal interests of video game operators participated in the cases previously considered, likely because no industry-specific trade association existed until 1994. The City of Warren court seemed especially unsatisfied with the examples presented to the record: the plaintiff made no specific arguments about the speech-related merits of electronic games but instead relied on previous acceptance of the genre as constitutionally protected without making a case to that effect when in fact no such precedent existed. In a similar vein, in Marshfield Family Skateland v. Town of Marshfield, another case seeking First Amendment protection for arcades, the plaintiffs submitted video excerpts of five games: Ms. Pac-Man, Tron, Donkey Kong, Zaxxon, and Kangaroo.33 While these games were popular, they may not have been the most relevant examples, either for the purpose of demonstrating electronic games capacity to transmit ideas or information, or more broadly as a form of expression. Caswell relied only on a description of the rudimentary plot of Space Invaders. Were video games of the early 1980s simply technologically advanced pinball machines34 or were the courts simply not presented with examples that effectively showcase their expressive capacity?

TheStateofElectronicGamesintheEarly1980s
The majority of games in the 1980s may well have lacked a sufficient informational element to meet the standard applied in Americas Best.35 Like many other emerging forms of media, early electronic games were highly derivative of other formats that shared cultural or structural characteristics.36 Because electronic games are seen as interactive entertainment, they mimicked previous examples in that genre, especially sports. By mimicking sports, game developers were making a safe bet: they could adapt proven game mechanics to a digital interface and potential players would be already familiar with the premise. Thus, from 1978 to 1979, Atari, then a major player in the industry, released Atari Baseball, Atari Basketball, Atari Football, Atari Mini Golf, and Atari Soccer.37 Other games submitted to court record may not be particularly compelling arguments for the informational capacity of electronic games. Kangaroo and Donkey Kong are platform games where players are given a simple premise: the playercontrolled title character must avoid danger to proceed up a series of platforms to rescue an innocent. In Zaxxon and Space Invaders, players pilot a spaceship and fire at oncoming enemies for points. It could be fairly argued that the relationship between Space Invaders and its plot was roughly equivalent to the relationship between chess and its medieval settingaesthetically helpful for setting the mood for the player, but unrelated to any message or information conveyed by gameplay.
Rothner v. City of Chicago. U.S. Court of Appeals, 7th Circuit 929 F. 2d 297. (1991) Marshfield Family Skateland v. Town of Marshfield. Supreme Judicial Court of Mass., 389 Mass. 436. (1983). 34 Marshfield, 444. 35 As has been and will soon be even more apparent, I do not find this standard compelling in theory or by precedent. However, it was the relevant standard embraced by the courts judging electronic games in this era. 36 Cf. Walter Benjamins seminal essay The Work of Art in the Age of Mechanical Reproduction, and his discussion of early film enthusiasts desire for the classics of literature to be made into films likened to heroes crowding the gates. 37 The Killer List of Video Games. Coin-Op Museum Index: A. KLOV.com. Accessed February 15th, 2010.
33 32

Rousse

It seems reasonable to speculate that the majority of games released in this period would not meet the informational capacity standard applied by the respective courts. Examining only the material submitted to the record, the courts may have come to the correct conclusion using this standard, but more general claims as to the unsuitability of electronic games to convey informational content are too broad. The material submitted in the respective cases did not provide an adequate overview of the expressive range of electronic games at the time. Yet, as cases are inevitably condensed and simplified when cited for precedent,38 subsequent courts might be led to believe that the entire medium of electronic games was unworthy of First Amendment protection. As we have seen, narrowly focused judgments on a small subsection of games have had far-reaching influence in the decisions of other courts regarding the entire format. If other games had been examined more carefully and with a more open mind towards their expressive potential, courts might have been persuaded that electronic games warranted First Amendment protection much sooner. Identifying media that have achieved expressive capacity is vital to protecting the First Amendment sphere, but our legal history does not have an excellent track record in this regard.

MissileCommandasExampleofProtectedSpeech
While the informational capacity standard may not be most appropriate for determining what media constitute speech, it was widely cited in the previously discussed cases of free speech regulation. This segment intends to demonstrate that electronic games were capable of meeting even the narrow definition of speech embodied by the information standard, but that the record presented to the court by the legal representatives of small business owners and the ability of the court to interpret a novel medium led to incorrect decisions regarding whether electronic games should be considered speech.. Assuming for the sake of argument that the capacity for communicating ideas or information is a valid test of speech, Missile Command, a popular game released in 1980 by Atari and designed by Dave Theurer, meets the criteria applied by the courts to the cases we have examined. In Missile Command, the player is presented with a flat landscape featuring six representations of cities and control of a missile battery. Via a trackball and button, the player fires counter-missiles from the missile battery to defend his or her cities against an increasing onslaught of missiles raining down from above. By striking these oncoming missiles, the player defends his or her missile battery and cities. Like many arcade games of the time, there is no winning condition as long as the player is good enough, or continues to put quarters into the machine, the game continues. It is difficult to evaluate Missile Command, released in 1980, outside of the context of the Cold War and fears of nuclear warfare and subsequent Mutually Assured Destruction. The original arcade release provided the player with no background story to situate the gameplay the player was simply faced with a barrage of oncoming missiles, unprotected cities, and the ability to fire counter-missiles. Protecting cities and destroying incoming missiles increased the players score. During the height of the Cold War, it seems likely that many players viewed this situation not as fantasy but as a (possibly imminent) hypothetical or an extrapolation of the tense nuclear standoff in the real world. Some oncoming missiles even mimicked the capabilities of real warheads. As the game progresses, lines representing missiles split during their descent, with each new line capable of damaging the players cities and silo. This mimics the capability of multiple independently targetable reentry vehicles (MIRVs), a variety of nuclear warhead capable of striking multiple targets in a single missile developed during the Cold War.39 In Missile Command, the trail of missiles splitting into multiple paths increases the challenge for the players and invokes the threat of the hundreds of MIRVs pointed at the United States by the Soviet Union at the time of the games release.
As we see here in the initial establishment of the informational capacity standard for free speech as a valid means to judge electronic games in the first place without a clear point of origin or coherent argument for its implementation. 39 Lawrence Buchonnet, MIRV: A BRIEF HISTORY OF MINUTEMAN and MULTIPLE REENTRY VEHICLES (Lawrence Livermore Laboratory, February 1976), http://www.gwu.edu/~nsarchiv/nsa/NC/mirv/mirv.html. Rousse 7
38

The fatalism of never-ending levels eventually resulting in the players defeat at the hands of an unending array of enemies built into the structure of most games of this period takes on a special poignancy in the context of nuclear war. In Missile Command, there is a never ending and steadily increasing number of incoming missiles and a finite supply of counter-missiles. Players must make tactical decisions when counter-missiles run low. If there are too few missiles to protect all of the cities, which would best be sacrificed? Which could be better defended? In order to play the game effectively, the player must make these choices- a grim task given the subject matter of the game. Ultimately, the players decisions are somewhat moot. No one wins in Missile Command, and this information is transferred to the player through the process of playing the game. The player likely applies the lesson to his or her experience of the real world as well: no one wins in nuclear warfare, either. This point is further driven home by the unusual screen shown at the end of a players game, when all of his cities have been destroyed. Rather than the typical GAME OVER common to most games of the period, Missile Command offers a stark black background with the text THE END.

Conclusion
The message of the game may be difficult to interpret to those who have not engaged in the material and are not familiar with the techniques of the medium, but so is much of expression protected by the First Amendment. Informational elements of many forms of protected high art with formal and abstract themes are far more obscure than that of Missile Command, further highlighting the weakness of the informational standard. If electronic games are held to the same standard as other media, information was conveyed by a video game to its players in 1980 at the latest, before any of the court cases previously examined, and thus should have been addressed as speech under the First Amendment. Furthermore, this information commented on a major political and cultural issue of its time period Missile Command shows that video games, even in the infancy of their development, could respond to contemporary culture. Compared to the mediums later development, Missile Command is an extremely rudimentary electronic game. The landscape never changes, the graphics are simple (the missiles are simply straight lines) and there is virtually no plot. There is simply the premise of the game: defending cities from a nuclear attack. There are certainly other expressive elements present in video games of this era. Many who have played Space Invaders will likely recognize that while it may not have much of a plot or feature cultural commentary, the pulsing crescendo of its soundtrack and the increasing speed of the alien menace are capable of conveying an emotive message of increasing anxiety. If we are to accept that a video game conveyed a message in 1980, then Winters v. New York40 suggests that we should not attempt to draw the line between entertainment and information within a medium capable of expression. Thus, the courts in Caswell, Americas Best, City of Warren, and Marshfield, using even the restrictive standard of informational capacity, are mistaken: electronic games, even in their early years when they were far from the contemporary products that have resulted from the rapid and unceasing development over the last three decades, should have received protection as speech under the First Amendment. An examination of the early case history of electronic game censorship does more than provide a context for Schwarzenegger v. EMA: it demonstrates that our current system of First Amendment protections is inadequate. Because there has not been a clear Supreme Court precedent that establishes a workable definition for speech, and thus establishes a workable definition for what media should be granted the protection of strict scrutiny, lower courts may classify new media as speech according to their own schema, which may contain significant flaws. If the First Amendment only protects media once they have become well-respected by mainstream culture and the narrow subset of mostly older and less technically savvy, it fails to deliver on its anti-majoritarian promise.

40

Winters, 65. 8

Rousse

References
COURT CASES Americas Best Family Showplace Corporation v. City of New York Department of Buildings. 536 F. Supp. 170 (E.D.N.Y. 1982) Caswell v. Licensing Commission for Brockton, 444 NE 2d 922 Mass: Supreme Judicial Court, Plymouth, 1983 City of Warren v. Walker. Court of Appeals of Michigan, 135 Mich. App. 267, 1984 Doran v. Salem Inn, Inc. 422 U.S. 922 (1975) Erznoznik v. Jacksonville, 422 U.S. 205 (1975) Jacobellis v. Ohio. 378 U.S. 184. (1964) Joseph Burstyn, Inc. v Wilson. 343 U.S. 495, 72 S. Ct. 777. (1952) Marshfield Family Skateland v. Town of Marshfield. Supreme Judicial Court of Mass., 389 Mass. 436. (1983). R.A.V. v. City of St. Paul, 505 U.S. 377, 395. (1992) Rothner v. City of Chicago. U.S. Court of Appeals, 7th Circuit 929 F. 2d 297. (1991) Schacht v. United States, 398 U.S. 58. (1970) Schad v. Borough of Mt. Ephraim. 452 U.S. 61. (1981) Southeastern Promotions, LTd. V. Conrad. 420 U.S. 546 (1979) United States v. Carolene Products, 304 U.S. 144, (1938). Winters v. New York. 333 U.S. 507 (1948) OTHER WORKS Buchonnet, Lawrence. MIRV: A BRIEF HISTORY OF MINUTEMAN and MULTIPLE REENTRY VEHICLES (Lawrence Livermore Laboratory, February 1976), http://www.gwu.edu/~nsarchiv/nsa/NC/mirv/mirv.html. The Killer List of Video Games. Coin-Op Museum Index: A. KLOV.com. Accessed February 15th, 2010. Winkler, Adam. Fatal in Theory and Strict in Fact: An Empirical Analysis of Strict Scrutiny in the Federal Courts. Vanderbilt Law Review, Vol. 59, p. 793, 2006

Rousse

The Game Behind the Video Game: Business, Regulation, and Society in the Gaming Industry New Brunswick, New Jersey, USA, April 8-9, 2011

DESIGNING FOR DIVERSITY: CAN REGULATION PROMOTE DIVERSE EXPERIENCES IN GAME WORLDS?

Dr Melissa de Zwart, Law School, University of Adelaide Adelaide, Australia Kate Roth, Monash University Melbourne, Australia ABSTRACT The creation of UCC [user created content] usually boosts the availability and diversity of local content in diverse languages. With lower entry barriers downstream and increased demand for content and lowered entry barriers upstream, the creation of content and overall cultural wealth could be positively influenced and the identification of artists facilitated. (OECD Working Party on the Information Economy, Participative Web: UserCreated Content, 12 April 2007, DSTI/ICCP/IE(2006)7/FINAL, 41) This paper will consider the regulatory factors that impact upon game diversity. Although there is an increasing recognition that cultural diversity is desirable as a function of democracy and participation in the public sphere, (Kerr, 2010) game worlds still largely reflect their cultures of origin. There have been some deliberate attempts to create merged environments, which reflect elements of key cultural attributes of their target markets, such as the highly successful Final Fantasy franchise (Consalvo, 2006). However, these environments are still the exception. This paper will consider if increased regulation will encourage or stifle diversity. It will analyse and discuss examples such as the swastika controversy in Call of Duty, Black Ops; protests over the alleged appearance of a Rising Sun flag in the Chinese MMOG Fantasy Westward Journey (Chan 2009); and attempts to both encourage and control User Generated Content (UGC). It will consider the effect of regulatory measures such as the Council of Europe/ ISFE Human Rights Guidelines for Online Games Providers (2008). This paper recognizes that one of the key values of MMOGs is their capacity to deliver a diversity of experiences and it is essential that this diversity is respected. Whilst input of UGC goes some way to embody and reflect diversity and user interest, there is potential for this to be stifled by internal and external regulation. Thus governance and regulatory intervention from domestic governments (and potentially through multilateral treaties and other regulatory instruments) must respect and support content diversity in order to ensure that MMOGs are able to continue to deliver valued user experiences (de Zwart 2009, de Zwart 2010). KEYWORDS: online games, diversity, user generated content, regulation

de Zwart and Roth

INTRODUCTION

The sheer size and output of the videogame industry, encompassing console, PC and online games and MMOGs, would suggest that sufficient diversity exists to cater to different user expectations, needs and tastes. However, the expectation of global delivery of game titles suggests that, rather than reflecting and fostering diversity, such games may reflect a single homogenous norm. Further, might this trend suggest a drive towards the lowest common denominator, such as was reflected in US TVI policy of Least Objectionable Programming up to the 1970s (Nelson 2007), content that is least offensive to the widest number of people, and hence, largely bland? Thus commercial imperatives for game and MMOG developers would suggest adherence to a set of production values that appeal to the broadest possible audience. Despite this, audience expectations may differ significantly between regions, in terms of language, culture and content. For example, Australia currently has no 18+ classification for computer games, so any violent or sexual content must be accommodated at the lower level of classification (suitable for children of 15 and above). There are also issues of cultural norms, such as racial stereotypes, portrayal of children, reference to historical and contemporary events, and more generally, cultural references and contexts. Thus this paper asks if regulation could have any role in fostering diversity and if in fact this would be desirable. Any regulatory environment requires careful management given the impact on user activities. In addition, as noted above in the OECD paper on User-Created Content (2007) the input of user created content and user-experience may increase the accessibility of local content and certainly has appeared to improve the commercial success of certain games. However, user created content brings with it risks of its own, uncontrolled copyright infringement, unsuitable content, defamation and content which is inconsistent with the developers vision for the game. Therefore, this paper will consider the need for diversity in games and how that may best be fostered, through government regulation, managerial techniques, the way users interact with the game or other regulatory frameworks. It will also focus upon the role that UGC may serve in this context.

DIVERSITY While I fully support diversity in real life, I dont necessarily find the need for it in my fantasy life. (Ozmodan, post to MMORPG.com, 11 December 2009)

What is meant by diversity in the game context? Clearly gamers themselves have differing views on the value of diversity. The discussion on MMORPG.com cited above involved in-depth discussion of the need for greater representation of women, people of any colour other than white, gays, and male and female avatars in games, with little agreement about the purpose that might serve. All of this ignored the fact that the games themselves are steeped in particular cultural norms and tropes. For example, does the existence of blue-skinned avatars legitimately encompass and represent different races? How do you portray actual historical events without glorifying acts of racism and xenophobia? We have identified several contexts for the consideration of diversity: Industry diversity: There have been some studies undertaken regarding the lack of diversity reflected within the game industry itself. The International Game Developers Association is undertaking a project on diversity to encourage underrepresented groups, i.e. anyone other than young, white males, into the game industry (IGDA 2005). Cultural diversity: Cultural diversity reflects and respects the manifold ways in which the cultures of groups and societies find expression (UNESCO Convention
2 de Zwart and Roth

on the Protection and Promotion of Cultural Diversity of Cultural Expressions) (Burri et al 2009). As Kerr (2010) has observed, the concept is far from uncontested, but that at least it is seen as a normative ideal for many and linked for some to a properly functioning public sphere and democracy. Graber (2010) states that games are recognised as a form of cultural expression in the digital environment, reflecting particular forms of language and content, and should be subject to similar forms of protection as other forms of cultural expression, such as art and literature: Digital games are undoubtedly a form of creative expression similar to other works of art, including, for example, verbal, musical, graphic, sculptural or audiovisual art. It seems reasonable to argue that as digital manifestations of art and story-telling, games should recognised as cultural productions. Several European countries now provide financial assistance to the production of games with sufficient cultural content (Graber 2010). However, this is occurring in the face of the fact that game production, at least at the higher production end of the spectrum, is concentrated in a small number of Japanese and US production houses, with the market divided into three regions: America, Western Europe and the Asia-Pacific (Kerr 2010, Johns 2006). In an age of increasing production costs, such companies are also increasingly looking for safe marketable content that will justify production and development costs. A Hollywood blockbuster movie such as Avatar is a safer commercial investment than an entirely new title. The intellectual property rights will also be cheaper to secure due to existing arrangements with the film studio and advantageous tie-in and marketing rights. Diversity of user experience: Users choose games and online worlds that suit their skills and interests. These may also include social motivations, i.e. to play with their friends. Within games there is frequently significant choice regarding player experience, character choice and customisation, PvP mode, etc. Thus one players experience of World of Warcraft may be very different from anothers. However, casual games currently provide little user customisation or choice. Diversity of player type / motivation: This has been discussed at length elsewhere. Notably Bartle divided players into achievers, explorers, socialisers and killers (1996). Other authors such as Yee (2007) have continued this work, identifying expanded player categories and motivations. Recently, Kallio, Mayra and Kaipainen (2010) have developed a classification of gamers based on gamer mentalities, derived from their belief that digital gaming [is] a cultural practice that is rapidly becoming a part of everyones everyday life everywhere in the world. Their analysis reflects the fact that people may game in different contexts and for different reasons and by identifying the gaming style rather than the gamer, this may assist a larger number of people in identifying themselves as gamers, without their practices being isolated from their broader social experiences. Kerr also notes that player preferences vary by gender, market and platform, for example, FPS games dominate the Xbox market in the US, but do not sell well in the Japanese market (2010). Game environment diversity: Games are now accessible on a broad variety of platforms, including PCs, both at home and more commonly in places such as South Korea in group environments, such as PC Baangs, console games, MMOGs, and increasingly, casual games such as Farmville on Facebook Plants vs Zombies on iPad, and other mobile devices. Diversity also exists in the types of games, FPS, action, role-playing games, quests, sport, childrens games etc.

Thus diversity can have different meanings according to context. However, we argue that diversity is desirable in all of these contexts, to reflect cultural backgrounds, user needs and
3 de Zwart and Roth

experiential levels. It would not be desirable to allow trade and regulation policy to create a situation of global game homogeneity. Bilateral agreements, such as the US-Australia Free Trade Agreement, and multilateral agreements such as the General Agreement on Tariffs and Trade (GATT), inextricably link intellectual property (and hence cultural products) and trade. Such trade agreements may provide the avenue for economic domination by a limited number of global players in the game industry, by removing the protection of local content benefits granted to cultural goods and requiring them to compete without the benefit of subsidies or tariffs in a global marketplace. Thus cultural values are placed second to removing barriers to trade in goods. Whether it is possible to mandate diversity is another matter. The Council of Europe has recommended that in developing and publishing games, the provider (defined as designers and publishers) should consider how such games may impact on human dignity with a particular emphasis on children (although it is worth noting that this is not necessarily a target group of the games and MMOs discussed below). The guidelines suggest that, amongst other things, regard should be had to: Content which conveys messages of aggressive nationalism, ethnocentrism, xenophobia, racism or intolerance in general including when such messages are concealed. (Council of Europe, 2008). Again, it is not clear how this translates to imaginary realms. In any event, there is greater emphasis upon protecting the children, than fostering cultural content or diversity. Thus existing regulation does not significantly foster diversity in game content. 3 GAMING CASE STUDIES

The section will present four case studies that explore different aspects of game content that give rise to issues of cultural diversity. 3.1 Final Fantasy Final Fantasy is a highly successful game franchise developed by Square Enix, a Japanese company created in 2003, following the merger of Square (originally founded in Tokyo in 1985) and Enix, another Japanese company. Final Fantasy (FF) installments were originally defined as independent stories with various settings and leading roles, however, there are recurring elements of FF including: plot themes, character names, and game mechanics. Plots center on a group of heroes battling a great evil while exploring the characters' internal struggles and relationships. Character names are often derived from the history, languages, and mythologies of cultures worldwide (Final Fantasy Collection, 2010). Square Enix has sold over 62 million units of the various Final Fantasy games worldwide. It now has subsidiary corporations in North America, the UK and China. As Consalvo (2006) explains the most important reason that Square Enix has been so successful in the American market (and other foreign locales) is the sophisticated and extensive localization used in translating its games for the USA and other markets. This localisation involves far more than merely translating the dialogue. It involves the musical score, text and images that appear in the game, resizing text and pop up boxes which contain text, changing local idioms and incorporating changes to take account of local content regulation laws or guidelines and user expectations (Carlson and Corliss 2010). Carlson and Corliss also remind us that these user expectations are heavily mediated through the understandings of local marketers and localisers. Changes can also be made to the underlying game engine to speed up or slow down certain elements of play and to comply with local hardware requirements. For example, one of the most popular iterations of the game, Final Fantasy XIII, was released in Japan in 2009 and in North America, Europe and Australia in 2010. Featuring the glamorous warrior Lightning as the main protagonist, FFXIII continued the tradition of having a racially mixed and ambiguous band of reluctant heroes, such as the group described by Consalvo (2006)
4 de Zwart and Roth

with respect to FFX (released in 2001 in Japan and 2002 in the US). Lightning is 171cm tall, with light pink hair, and pale skin. She is only 21 years old, but a seasoned warrior. She is clad in a highly stylized outfit, which emphasizes her long legs, and wields the Blaze Edge, a combination of a gun and a sword. Her character, as with all of the others, is voiced by different actors in the Japanese and the English versions. She is accompanied by a band of male and female characters of diverse appearance, including Oerba Yun Fang, who is of vaguely Indian appearance sporting a stylized sari, although her accent is Australian, and Sazh Katzroy, who has dark skin and a large afro, which conceals a baby Chocobo (a flightless bird featured throughout the Final Fantasy series). Critics were divided by the game, generally loving its visual richness and the relationship between the key characters, however, there was significant criticism regarding the linear nature of the game, which limited side adventures and exploration of the environment. In an interview, the game director Motomu Toriyama responded: We think many reviewers are looking at Final Fantasy XIII from a western point of view. When you look at most Western RPGs, they just dump you in a big open world, and let you do whatever you like... [It] becomes very difficult to tell a compelling story when you're given that much freedom (Ingham 2010). One further point worthy of note was the translation of the musical soundtrack to the game. In the US version of the game the original theme song Kimi ga Iru Kara by Sayuri Sugawara was replaced by My Hands by Leona Lewis. Naturally fans were divided about the use of a nonoriginal song for the soundtrack and one that did not have the same cultural flavour as previous theme songs. Many expressed the view that Lewis, as a former winner of the X Factor, was too mainstream for a computer game, and that this was an instance of the developer selling out. In another interview, producer Yoshinori Kitase and director Motomu Toriyama explained: We felt that with a song that's sung in a language that's understandable to North American or European users would bring the game closer to the player and depart from the idea that Final Fantasy is a game that comes from overseas. Further, they claimed: Overall, it would tighten the relationship between the player and the game, so that's why we decided to go with an English theme song (Jackson 2010). On 30 September 2010 the latest instalment of the franchise Final Fantasy XIV was released as an online game. The next day online discussion groups flourished with sentiments expressing gamers disappointment with the styling, levelling, game-play and lack of direction. The critics were similarly scathing, with ratings of 49 out of 100 from Metacritic and 4 out of 10 from Gamespot. Subsequent activity by the developers has seen a whole new team put in place and announcements of re-release after re-release. The development of an online game is a change in direction from the franchises typical projects, and fans seem to be wedded to the video game platform. However, with a series of upgrades to the online version attitudes may change. This development is in line with the range of other spin-off projects developed from the Final Fantasy franchise branching into motion pictures, anime, printed media, and other merchandise (Final Fantasy Collection, 2010). Notwithstanding the past successes of the game, Final Fantasy XIV online is still plagued with problems and, since commencing this paper, a further three major disruptions to the environment have prevented significant game play. The Final Fantasy experience would suggest that game developers themselves strive to achieve glocalisation for their games (Consalvo 2006) in order to ensure commercial success. Although as Carlson and Corliss (2010) point out, this appears to run directly counter to the current popularity of specifically Japanese games, anime and manga in the US, Europe and Australia. 3.2 Call of Duty, Black Ops Call of Duty, Black Ops (CoD-BO) is the seventh instalment in the successful Call of Duty franchise and is available on multiple platforms, including Xbox 360, PS3, Wii and DS. It is set in
5 de Zwart and Roth

the Cold War environment of the 1960s, and features a number of CIA initiated black ops behind enemy lines. Notably, one of these is a failed assassination of Fidel Castro, which generated significant controversy in Cuba (BBC 2010). However, the event that generated the most controversy to do with the game was the request by an Xbox Live user to use a swastika as his online profile symbol. The request was met with outright refusal by Microsofts XBox Live Director of Policy and Enforcement, Stephen Toulouse, who tweeted: Yes, you will get banned using a swastika as your Black Ops emblem. I dont care if you just like the design. Following a barrage of criticism and Twitter-fuelled debate over the rights arising under the US First Amendment, Toulouse went on to blog about his decision as follows: The Xbox LIVE profile and in game content you create is accessible by everyone. You do not have the context inside of it to explain your long winded contrarian view that your pithy text that violates the Terms of Use or Code of Conduct is actually intended to change peoples minds about a commonly held understanding. Its not political correctness, its fundamental respect. If you think the swastika symbol should be re-evaluated by societies all over the Earth, I think thats great. Your Xbox LIVE profile or in game logo, which doesnt have the context to explain your goal, is not the right place to do that. (Toulouse, 2010) Journalists and newspapers carried the story and tried to generate arguments and controversy. Despite this, the majority of blogs and responses indicated support for Toulouses stance as well as understanding the level of censorship that Xbox Live placed on CoD-BO content (Guardian 2010; TG Daily 2010). Naturally, this was not the first time Call of Duty had attracted controversy. Call of Duty: World at War is set during World War II. It features extensive cut scenes, using actual archival footage. For example, the introductory sequence shows the Nazi swastika bleeding across Europe and the red sun bleeding across Asia, followed by a very rapid sequence of images showing airplanes, soldiers marching, the mass shooting of civilians, the bombing of Pearl Harbour and the US declaration of war (Baron, 2010). All of this occurs in one minute and twenty-four seconds and provides the perfect introduction to the player being transformed into Private Miller a soldier ready to do battle against the forces of evil. Baron argues that this provides the sense that history marches only in one direction and indeed legitimises the violence to be used by the player against the Germans and the Japanese for the rest of the game. Baron suggests that, rather than merely forming the prologue to the violence, archival footage might be used across a range of scenarios, dependent on the user's actions within the game-play, thereby, proposing other possible histories that might have occurred if certain battles had not been won. It is inevitable that games dealing with war will cause offence to some sector of the community. For example, the use of the swastika and other Nazi symbols is banned in Germany. However, it is also true that war games may be used to reinforce racial stereotypes and to alienate the other. Authenticity and realism can be used as a cover for killing Gooks and Towel Heads with impunity and to imbue certain races with stereotypical behaviours (Chan 2005). These examples indicate that different nationalities and cultural groups have varying limits upon freedom of content in gaming contexts. 3.3 Rising Sun Controversy Fantasy Westward Journey is a very popular Chinese MMOG, with a peak of over 25 million registered player accounts in 2006 at the time of the events discussed below. It is an adventure
6 de Zwart and Roth

game based on a classic of Chinese literature. In July 2006, a high level player was asked by the game administrators to change the name of his avatar, which was called Kill the Little Japs. He refused to do so and, as a consequence, the avatar was suspended from engagement with the game. He had also established a guild, called The Alliance to Resist Japan which boasted around 700 members, making it one of the largest in the game. The game administrators also informed the guild members they would be dissolving the guild (Chan 2009). On 7 July 2006, thousands of avatars congregated at an area in the game known as the Jianye city government office protesting these actions, and directed attention to a wall of the office, claiming that it depicted the Japanese Rising Sun flag. The alleged appearance of a Rising Sun flag in the game prompted other in-game protests and challenges of corporate interference, claiming that such icons resulted from a proposed Japanese take over of the game, through a sell out of NetEase (the game owners) to a Japanese company (Chan, 2009). Notably, this all occurred on the anniversary of a significant battle between the Chinese and Japanese in 1937. Chans recounting of the story canvasses the conspiracy theories behind the break up of the guild, such as a belief by the game providers that the particular guild was too big or too powerful and the claim that the actions were undertaken as the game owners were looking to sell and were conscious of minimising any potentially offensive content or activities in world. An alternative theory which Chan explores is that these actions were in fact manipulated to generate spontaneous nationalistic sentiments which would be adopted by gamers in-world to promote strong patriotic Chinese cultural identification in a significant and relevant online environment. Carlson and Corliss (2010) also explore the political elements of localisation which adopts a particular stance regarding what is acceptable in a particular zone and targets modifications accordingly, for example, to pre-empt the requirements of national ratings boards or by the use of a particular language or dialect as the dominant language in translations. For example, in Australia, it is a legal requirement under the National Classification Scheme that computer games be classified before sale. The highest rating which may currently be applied with respect to a computer game is MA 15+, meaning that the game must be suitable for players aged 15 and above. Thus content can be manipulated to serve multiple purposes, both to promote and deny diversity. 3.4 User Generated Content As Burri et al (2009) observe, although the empowerment of the user through engagement in cultural production through UGC creates the promise of better, more efficient and flexible accommodation of the goal of cultural diversity, it also carries with it challenges which may demand additional regulatory intervention. However, UGC brings with it at least a promise of greater user engagement and involvement in the process of production and consumption. The ability for UGC to be incorporated into game development and associated products (such as fan fiction, machinima, mods, add-ons, walkthroughs and web sites) provides more timely and more genuinely local responses to the game or the franchise than the developer or publisher can provide as it is unmediated by professional marketers. Publishers are aware of the value of this, and have generally assisted, supported or at least turned a blind eye to these activities, despite the fact that they are largely in breach of the EULA. However, they are of course concerned with UGC quality and content, the risk of infringing third party intellectual property rights, content regulation issues and potential damage to the brand through undesirable associations. Some games actively seek to co-opt the energy and appeal of UGC, such as Sonys Little Big Planet (LBP). However, LBP experienced significant difficulties when it provided a platform for user generated levels. A large number were found to contain material which infringed third party intellectual property rights and were therefore removed, aggravating users who had been led to believe that they had free reign over their creations, only to find them removed without warning (Graft 2008). In addition, many
7 de Zwart and Roth

users are unaware of the terms of the EULA which provide Sony with a right to commercially exploit UGC material. Guoshong (2009) argues that players who engage in UGC creation consume contents for fulfilling their information, entertainment, and mood management needs; they participate through interacting with the content, as well as with other users for enhancing social connections and virtual communities; and they produce their own contents for self-expression and self-actualization. This would suggest that users engage with games and produce UGC for largely non-commercial reasons. However, it does not follow that users realise that their labour may result in a commercial product with no financial benefit flowing back to them. These issues will need to be settled before UGC can be advocated as a viable solution to a lack of local content. 4 CONCLUSION

Games should be as diverse as their players: Final Fantasy builds on a successful franchise of activities from the last 27 years, pitting bands of youthful warriors against the forces of evil; Call of Duty uses historical imagery to set up in-game story lines, drawing upon real life footage and events to create new gaming experiences; and Fantasy Westward Journey appears to serve a role in bringing traditional folklore to life and in promoting modern-day nationalistic sentiments. The events in the games discussed in this paper demonstrate the commercial imperatives for customising games for local markets, noting the limiting frameworks of national regulation of content, the political potential of games and the effectiveness and limitations of user generated content in facilitating development of local flavour. This paper concludes that rather than mandating and regulating for diverse content, regulators should allow games to evolve to meet user demand. Intervention may only be necessary to support funding for niche domestic cultural reasons or to accommodate smaller ethnic groups. These activities should not be impeded by agreements forged to regulate interstate commerce. UGC provides an opportunity for the development and continuation of games and the cultures they develop. In online gaming, the choices of subscribers to a particular environment provide the ultimate decision, by their choice to be and remain part of the environment. However, for UGC to provide a significant role in localisation and diversity, intellectual property issues need to be resolved. The intentions of game developers and providers need to be reflected in the EULA, rather than there being a mismatch between the legal constitution of the game and its underlying logic and intent. EULAs should be drafted for a specific environment, rather than bolted on as a one size fits all by the legal department to cover all contingencies, relevant or not. The development of games and their customisation for local markets and tastes provides an interesting basis for consideration of elements of diversity. Indeed, as we have seen, in attempting to ensure the national popularity of a title, different games may adjust story lines and outcomes in releases across the globe, often based on a stereotypical understanding of that culture. These adjustments, while based on tested as well as perceived realities, result in alterations to a game environment only where it is seen as financially efficient. If the adjustments are too expensive for a particular title, it may not be released in that region or released with basic customisation only. Some would argue this perpetuates stereotypes and others see this as an evolution in gaming, developing significant meanings, interpretations, in turn contributing to an ongoing evolution of further development and understandings. For example, Malaby (2010) has recently argued that one of the key attributes of games is 'semiotic contingency' or the unpredictable and changing nature of meanings that can be attached to game related acts or consequences. He observes: The complex contingency of today's large-scale online games has powerful effects on meaning, conceived here as always arising from the meeting point of existing, shared
8 de Zwart and Roth

interpretive frameworks and unique, contingent circumstances. Additionally, as games themselves show most powerfully, the shared engagement of contingency is a powerful means for the development of trust and belonging. The very nature of development and evolution in games responds to consumer and regulator requirements and provides examples and experiences of diversity. The protection of these frameworks ensures evolution and enables diverse experiences in such environments. It is important that any regulation of games seeks to facilitate and support diversity, not by mandating particular cultural norms but by supporting development of games that reflect cultural values. In this regard, trade related aspects of intellectual property enforcement should be carefully monitored to respect and protect local content production. Work should also be undertaken by game providers to ensure they have EULAs that are consistent with their expectations and encouragement of UGC, as appropriate. Only then will diversity flourish for all gamers regardless of their field of endeavour.

REFERENCES AND CITATIONS Bartle, R. (1996) Hearts, clubs, diamonds, spades: Players who suit MUDs. http://www.mud.co.uk/richard/hcds.htm. Baron, J. (2010) Digital Historicism: Archival Footage, Digital Interface, and Historiographic Effects in Call of Duty: World at War. Eludamos. Journal for Computer Game Culture. 2010; 4 (2), 303-314. BBC (2010) Cuba says Castro video game tries to legitimise murder, 11 November 2010, http://www.bbc.co.uk/news/world-latin-america-11731120. Burri, M., Graber, C.B. and Steiner, T. (2009) The protection and promotion of cultural diversity in a digital networked environment: Mapping possible advances to coherence. NCCR Trade Regulation, Working Paper No 2009/36. Burri-Nenova, M. (2010) User created content in virtual worlds and cultural diversity in Graber, C.B. and Burri-Nenova, M. (eds.) Governance of Digital Game Environments and Cultural Diversity, Edward Elgar, Cheltenham, UK. Carlson, R. and Corliss, J. (2010) Imagined Commodities: Video Game Localization and Mythologies of Cultural Difference. Games and Culture. 6. pp 61 - 82 Chan, D. (2005) Playing with Race: The Ethics of Racialized Representations in E-Games International Review of Information Ethics 4, 24- 30. Chan, D. (2009) Beyond the Great Firewall The Case of In-Game Protests in China. Hjorth, L. and Chan, D. (ed.) Gaming Cultures and Place in Asia-Pacific. Routledge, New York. Classification Policy Website. Attorney General Website http://www.ag.gov.au/www/agd/agd.nsf/Page/Classification_policy. Accessed 6 March 2011. Consalvo, M. (2006) Console video games and global corporations: Creating a hybrid culture. New Media Society. 8, 117 -137. Council of Europe. (2008) Human Rights Guidelines for Online Games Providers. de Zwart, M. (2009) Piracy vs Control: Models of Virtual World Governance and Their Impact on Player and User Experience. Journal of Virtual Worlds Research.2.3. de Zwart, M. (2010) Contractual Communities: Effective Governance of Virtual Worlds. University of New South Wales Law Journal. 33.2. 605-627. Final Fantasy Collection. http://en.wikipedia.org/wiki/Book:Final_Fantasy_series
9

de Zwart and Roth

Gamespot Website; http://au.gamespot.com/pc/rpg/finalfantasy14/review.html?tag=topslot;thumb;1 Accessed 20 February 2011. Graber, C.B. (2010) State Aid for Digital Games and Cultural Diversity: A Critical Reflection in Light of EU and WTO Law in Graber, C.B. and Burri-Nenova, M. (eds.) Governance of Digital Game Environments and Cultural Diversity, Edward Elgar, Cheltenham, UK. Graft, K. (2008) Sony: Be Mindful of LittleBigPlanet Content Edge, 13 November 2008, http://www.next-gen.biz/news/sony-be-mindful-littlebigplanet-content. Guosong, S. (2009) Understanding the appeal of user-generated media: a uses and gratification perspective, Internet Research, Vol. 19 Iss: 1, pp.7 25 Ingham, T. (2010) Final Fantasy XIII boss responds to review scores Computer and Video Games.com, 16 February 2010, http://www.computerandvideogames.com/235030/news/final-fantasy-xiii-bossresponds-to-review-scores/. International Game Developers Association (2005) Game Developer Demographics: An Exploration of Workforce Diversity, http://archives.igda.org/diversity/report.php. Jackson, M. (2010) Leona Lewis brings Final Fantasy XIII 'closer to the West', 15 February 2010, http://www.computerandvideogames.com/234805/news/leona-lewis-brings-finalfantasy-xiii-closer-to-the-west/. Johns, J. (2006) Video games production networks: value capture, power relations and embeddedness. Journal of Economic Geography 6, 151-180. Kallio, K., Mayra, F. and Kaipainen, K. (2010) At Least Nine Ways to Play: Approaching Gamer Mentalities Games and Culture, DOI:10.1177/1555412010391089. Kerr, A. (2010) Beyond billiard balls: transnational flows, cultural diversity and digital games in Graber, C.B. and Burri-Nenova, M. (eds.) Governance of Digital Game Environments and Cultural Diversity, Edward Elgar, Cheltenham, UK. Luttrell, M. (2010). Call of Duty spurs strong anti-swastika message on Xbox. TG Daily. 22 November 2010. http://www.tgdaily.com/games-and-entertainment-brief/52655-call-ofduty-spurs-strong-anti-swastika-message-on-xbox. Malaby, T. '(2007) Beyond Play: A New Approach to Games', Games and Culture, 95- 113 Metacritic Website; http://www.metacritic.com/game/pc/final-fantasy-xiv-online Accessed 20 February 2011. Nelson, R. (2007) HBO Premium: Channelling distinction through TVIII. New Review of Film and Television Studies, 5: 1, 25-40 OECD Working Party on the Information Economy, Participative Web: User-Created Content, 12 April 2007, DSTI/ICCP/IE(2006)7/FINAL, 41 Stuart, K. (2010) Call of Duty: Black Ops is Microsoft right to ban the swastika? The Guardian. 23 November 2010. http://www.guardian.co.uk/technology/gamesblog/2010/nov/23/call-of-duty-swastikacontroversy. TG Daily (2010) http://www.tgdaily.com/games-and-entertainment-brief/52655-call-of-dutyspurs-strong-anti-swastika-message-on-xbox The Guardian (2010) http://www.guardian.co.uk/technology/gamesblog/2010/nov/23/call-ofduty-swastika-controversy Toulouse Twitter account; http://twitter.com/Stepto/status/6196258494484480 Toulouse, S. (2010) StepTo.com. Context. Or, no you dont get to apply your Internet niche knowledge to me doing my job. :> 20 November 2010, http://www.stepto.com/Lists/Posts/Post.aspx?ID=635 Yee, N. (2007) Motivations for play in online games. CyberPsychology and Behaviour, 9: 772-775.
10 de Zwart and Roth

The Game Behind the Video Game: Business, Regulation, and Society in the Gaming Industry New Brunswick, New Jersey, USA, April 8-9, 2011

Productivity Games: Software Quality through Fun and Play


Ross Smith Microsoft Corporation One Microsoft Way, Redmond, Washington, USA

ABSTRACT There are a wide variety of definitions of quality, leaving a degree of freedom in how we pursue software quality. From ISO 9000 to Six Sigma, the diversity of viewpoints offers an opportunity to expand how we test software. While its important to understand the relationship between compatibility, performance, reliability, security, usability, and value its also critical to focus on customer satisfaction, which may vary, by user across these dimensions. Software test professionals face the law of diminishing returns when testing and re-testing software. Its important to incorporate principles of portfolio selection and risk diversification when evaluating where to invest resources. As the nature of the workforce shifts with incoming Gamer Generation employees, we can make use of successful game mechanics to help motivate employees to participate in work-related activities. The principles of trust, collaborative play, and fun allow novices to experiment with new strategies with the freedom to fail in game play that might not exist in the traditional workplace. Risktaking, supported by the structure of game mechanics, play, and organizational trust actually improves software quality by varying defect detection techniques. Productivity games games in the workplace can help enhance traditional workplace methodologies including effective communication, employee engagement, cost savings, and more. Employees desire many of the same things from the workplace that gamers demand from games This paper will take a look at productivity games successfully deployed at Microsoft over the last 7 years. Several thousand Microsoft employees in Windows, Office, and around the company, have played productivity games to help improve the quality of our software. We have learned much about what makes for a successful and an unsuccessful productivity game well beyond game design and more importantly, focusing on the tasks that game players can accomplish. KEYWORDS: Productivity Games, Software Test, Collaborative Play, Fun

INTRODUCTION

Kinect, the natural user interface add-on to XBox, recently became the fastest selling consumer electronics device in history.1 Around the world, hundreds of millions of people play games. The idea of play certainly resonates with humans, animals, and even computers themselves. We see games appearing everywhere these days except in the workplace. Merriam-Webster defines work as activity in

which one exerts strength or faculties to do or perform something 2and defines play as recreational activity; especially: the spontaneous activity of children3 Jane Mcgonigal, in her book, Reality is Broken, suggests that the opposite of play isn't work, its depression4 This paper suggests that perhaps its time to consider game play as a new way to work. The idea of
1 2

http://www.reviewsofelectronics.com/kinect-breaks-guiness-record/227632/ http://www.merriam-webster.com/dictionary/work 3 http://www.merriam-webster.com/dictionary/play 4 http://www.amazon.com/Reality-Broken-Games-Better-Change/dp/1594202850


Productivity Games: Improving Software Quality through Fun and Play 1

providing a structure - in the context of gaming, play, and game rules - for people to experiment with new approaches, new roles, or new ideas can lead to creative and innovative output, while keeping fun alive in the cold, cruel, competitive world of work.
2 SOFTWARE QUALITY 2.1 Definition

High quality is always an admirable goal for any software development project. To begin, we must decide on a definition of quality. Surveying several industries, associations, and experts, we find hundreds of definitions of quality, each of which seems reasonable and applicable. How to decide and pick a single definition to work towards? Here are a few examples: The perception of the degree to which the product or service meets customer expectations ISO 9000: Degree to which a set of inherent characteristics fulfills requirements. Six Sigma: Number of defects per million opportunities Philip B. Crosby: Conformance to requirements Joseph M. Juran: Fitness for use by the customer Gerald M. Weinberg: Value to some person Robert Pirsig: The result of care American Society for Quality: A subjective term for which each person has his or her own definition. In technical usage, quality can have two meanings: o The characteristics of a product or service that bear on its ability to satisfy stated or implied needs; o A product or service free of deficiencies." The answer is that we cant pick just one. All of these are perfectly viable. The variance in definition is a leading indicator of the importance of introducing variance and diversity in to our quality assurance processes.
2.2 Metrics and Measurement

Any discussion of metrics must start with clear identification of the organizations mission and goals. Without knowing the mission and goals around the quality-related efforts, any organization will have an extremely difficult time choosing metrics to provide an accurate picture of the progress. Software quality is important across several different dimensions. Figure 1 shows nine dimensions of software that are critical to user perception of quality.

Productivity Games: Improving Software Quality through Fun and Play

Figure 1 Nine dimensions of quality


2.3 Software Testing

Software testing is more effective when defects are discovered early. Whether the development process follows an agile or waterfall development model, the sooner in the process that defects can be uncovered, the more cost effective and impactful those fixes are to the product. One common fallacy is that the testing process can fully measure and quantify the degree of quality of the software. With the advent of the internet, multiple platforms, an explosion of usage scenarios, and broad diversity of user base and infrastructure, it is virtually impossible for a test team to understand everything. Early identification of an all-encompassing test matrix is untenable. Therefore, its critical to apply test effort and resources on a riskmitigation basis, rather than on a pure quality assessment scale.

Productivity Games: Improving Software Quality through Fun and Play

Figure 2 Diminishing Returns for Defect Discovery

2.4 Risk and Mitigation

James Reason, author of Managing the Risks of Organizational Accidents5, hypothesizes that most accidents can be traced to one or more of four levels of failure: Organizational influences, unsafe supervision, preconditions for unsafe acts, and the unsafe acts themselves. In this model, an organization's defences against failure are modelled as a series of barriers, with individual weaknesses in individual parts of the system, and are continually varying in size and position. The system as a whole produces failures when all individual barrier weaknesses align, permitting "a trajectory of accident opportunity", so that a hazard passes through all of the holes in all of the defenses, leading to a failure. The model includes, in the causal sequence of human failures that leads to an accident or an error, both active failures and latent failures.6 Software defects may occur in very similar circumstances organizational or project management related influences, unsafe supervision of the development process, preconditions for unsafe acts such as lack of knowledge of the end user scenario, and the unsafe acts (simple programming errors) themselves. As the complexity of the connected world continues to grow, it is just not possible to test quality in to the software. At best, system test can assess a reasonable level of quality across common scenarios. Even that, however, only scratches the surface of how software is used in real life. Therefore, its important to have as large a set of testers in as many configurations as makes reasonable sense. As Linus Torvalds describes Linus law in "The Cathedral and the Bazaar", "given enough eyeballs, all bugs are shallow"; or more formally: "Given a large enough beta-tester and co-developer base, almost every problem will be
5 6

http://www.amazon.com/Managing-Risks-Organizational-Accidents-Reason/dp/1840141050 http://en.wikipedia.org/wiki/Swiss_Cheese_model
4

Productivity Games: Improving Software Quality through Fun and Play

characterized quickly and the fix will be obvious to someone."7 So a broad set of pre-release users can help mitigate risk through broad test coverage. But, how can you motivate a broad set of users to perform tasks in their discretionary time to help test? Productivity games offer one possibility.
3 21ST CENTURY EMPLOYEES

Each generation imagines itself to be more intelligent than the one that went before it, and wiser than the one that comes after it. - George Orwell We are currently facing demographic and societal changes, economic landscape alterations, globalization, and the continuing rise of the knowledge worker that have led us to a workplace in the United States where members of four generations sit side-by-side, for the first time. Corporate America is as diverse as ever. An unprecedented number of international, multi-cultural workers from four generations Traditional, Baby Boomers, Gen Xers and Generation Y/Millennials are working alongside one another and bringing their own values, goals and communication approaches to the workplace. Such generational dynamics affect morale, productivity, recruitment and retention. Employers are facing immediate challenges when it comes to optimizing productivity, protecting profits and growing their businesses.8 The makeup of formal and informal - organizations has always mirrored that of society as a whole. As societal demographics shift, organizations have no choice but to consume the transformation. A failure to respond to the demographic changes of society will make it difficult, if not impossible, for the modern organization to meet employee needs and productively move forward.9 This will leave the organization unprepared to compete for talent or in the marketplace - eventually leading to its demise.
3.1 Computer Science Education

The quality of computer science degrees has grown dramatically over the last 40+ years. In the period from 1966 to 2001, the annual Bachelors degrees awarded in Computer Science grew from less than 100 to over 46,000 annually.10 The dot com boom fuelled continued growth, both in the numbers, and, given the demand, in the quality of education as well. As a result, college candidates entering the computer industry today have more education, more training, and more hands-on experience than candidates of even 10 or 20 years ago. Curriculum development has progressed dramatically as well particularly in software quality. While graduate level CS programs in the 1980s might have offered courses such as, Introduction to Basic Programming, today, Rutgers, for example, offers undergrad courses in Compiler Construction, Algorithms, and Cryptography. These changes in the quality of education obviously influence talent levels entering the workforce.

7 8 9

http://en.wikipedia.org/wiki/Linus%27s_Law
http://www.hrmreport.com/article/Engaging-the-Multi-generational-Workforce/

http://www.associatedcontent.com/article/1057728/changes_in_workplace_demographics_the.html?cat=17 10 Allen B. Tucker, Computer Science Handbook, p4.


Productivity Games: Improving Software Quality through Fun and Play 5

3.2 Global Workforce

The GDP growth dominance of emerging economies will continue to change the landscape and makeup of the workforce. The market for talent will continue to be highly diverse, networked, and distributed. 83% of the new global R&D facilities were in China and India (2004 - 2007). 91% of the increased R&D staff was in China and India.11 Indias middle class will swell by more than ten times its current size of 50 million to 583 million by 2025.12 Kenya leads the world in money-transfer by mobile phone.13 Mobile Phone Subscribers: 738.57M14 Globally, mobile data traffic will double every year through 2014, increasing 39 times between 2009 and 2014. Mobile data traffic will grow at a compound annual growth rate (CAGR) of 108 percent between 2009 and 2014, reaching 3.6 exabytes per month by 2014.15 The world is indeed getting flat. The ability to manage employees across social, cultural, geographic, technical, educational, generational, and political boundaries requires management 2.0 thinking. Productivity games are a part of the new management orthodoxy.
3.3 Gamer Generation

In every job that must be done, there is an element of fun. You find the fun andSNAP the jobs a game! Mary Poppins Games have been around for centuries. For example, the game of Go was played in ancient China in 500 B.C. and is for sale along most toy aisles and in game stores, as well as online.16 The royal game of Ur was played in the fourth century B.C. You can play it online today on the British Museum Web site17 . Evidence even seems to indicate that games were used to motivate workers to build the pyramids of Egypt.18 Although the current generation of electronic games has its roots in the 1950s and 1960s, with Spacewar and PONG, computer and game console developments over the last few decades have taken gaming to a new level. Perhaps it is unfair to attribute the start of a generation to a single game, but in 1981, Nintendo published Donkey Kong, which really was a defining moment in mainstream gaming. Typically, the majority of the programmers or engineers are relatively recent college graduates. Coincidentally, many of these same people grew up with video games as a significant part of their daily environment and were taught basic principles through their heavy interaction

(Booz & Company, 2009) (link) (MGI, May 2007) (link) 13 (Economist, April 2010) (link) 14 (Trendsniff, December 2009) (link) 15 (Cisco, February 2010) (link) 16 www.gnu.org/software/gnugo/free_go_software.html 17 www.thebritishmuseum.ac.uk/explore/families_and_children/ online_tours/games/the_royal_game_of_ur.aspx 18 The Practical Guide to Defect Prevention, Chapter 5 www.defectprevention.org
12

11

Productivity Games: Improving Software Quality through Fun and Play

with video games. The result is that they identify with and will respond to the idea of using games in their work activities. In their book The Kids Are Alright, John Beck and Mitchell Wade enumerate the basic principles that video games have taught this generation:19 If you get there first, you win. There is a limited set of tools, and it is certain that some combination will work. If you choose the right combination, the game will reward you. Trial and error is the best strategy and the fastest way to learn. Elders and their received wisdom cant help; they dont understand even the basics of this new world. You will confront surprises and difficulties that you are not prepared for. But the sum of those risks and dangers, by definition, cannot make the quest foolish. Once you collect the right objects (business plan, prototype, customers, maybe even profits), youll get an infusion of gold to tide you over. Although there may be momentary setbacks, overall the trend will be up. It is important to pay attention to these principles when designing games to motivate the work force to invest time in defect detection and defect prevention activities.
3.4 Organizational Trust

In a recent University of British Columbia report, economists found that trust in management is the most valued determinant of job satisfaction. They report that, in terms of job satisfaction, a small increase in trust of management equates to a 36 percent pay increase. Conversely, the researchers found that if that same amount of trust is lost, the decline in employee job satisfaction is like taking a 36 percent pay cut.20 There are many reasons why trust is important. Trust is a mechanism that people can deploy to deal with uncertainty. Dr. Mike Armour, in his book "Leadership and the Power of Trust" defines trust as the "complete confidence that a person or organization will consistently try to do what is right in every given situation." Employees who work together must rely on one another, either directly or indirectly, to be successful. Even if jobs are unrelated or people are not in physical proximity, the success of the firm depends on the ability of everyone to produce. Stephen M.R. Covey, in his book, The Speed of Trust, talks about the trust dividend high trust significantly improves communication, collaboration, execution, innovation, strategy, engagement, partnering, and relationships with all stakeholders. In your personal life, high trust significantly improves your excitement, energy, passion, creativity, and joy in your relationships with family, friends, and community. Obviously, the dividends are not just in increased speed and improved economics; they are also in greater enjoyment and better quality of life.21

The Kids Are Alright, John Beck and Mitchell Wade The HR Executives Role in Rebuilding Trust, Dennis S. Reina and Michelle L. Reina, http://www.hreonline.com/HRE/story.jsp?storyId=12160414 21 The Speed of Trust, Stephen M.R. Covey, p. 19
20

19

Productivity Games: Improving Software Quality through Fun and Play

Trust is a foundational element of all successful workplaces. The attributes of effective organizations all find their origins in trust. Gallup uses a Q12 survey to measure workgroup effectiveness and employee engagement22. Questions such as I have the opportunity to do what I do best every day or I know what is expected of me at work imply a level of trust not only by the employee for their manager or organization, but reciprocal as well.
3.5 Employee Engagement

Gallup does their annual survey and typically, the numbers have been very similar over the past few years. About 30% of employees are engaged at work, with about 55% who are passively disengaged, and about 15% who are actively disengaged. As generational changes in the workforce continue to evolve, the concerns over engagement continue to rise. Forty-five percent of Millennials worldwide say they use social networking Web sites at work, regardless of whether their organization or company prohibits their use.23 Two-thirds of teen and Gen Y Internet users use social networking sites; less than 10% of 55-63 year-olds do.24 The most successful employee engagement techniques, across this diverse set of behaviours, are challenging to identify and understand. It is very hard to find common themes to engage employees. Organizational leaders would like to believe that a strong mission and vision statement will suffice and while that helps its not necessarily all that is required.
4 PRODUCTIVITY GAMES 4.1 Introduction

From a Productivity Games viewpoint, the employee categorization and the organizational classification overlap in a way that can help identify whether or not a game will be successful in modifying behavior and having people play. Table 1 illustrates the areas where Productivity Games can be the most successful. Focusing either on expanding skills in role, or Organizational Citizenship Behaviors (OCBs) that require core skills are the best way to ensure the success of the game. There are several dimensions of citizenship behaviours. There is a five-factor model consisting of altruism, courtesy, conscientiousness, civic virtue, and sportsmanship. 25 Altruism is helping a specific work colleague with an organizationally relevant task or problem.

22 23

http://www.gallup.com/consulting/52/employee-engagement.aspx (Accenture, January 2010) (link) (Pew, January 2009) (link) http://en.wikipedia.org/wiki/Organizational_citizenship_behavior
8

24 25

Productivity Games: Improving Software Quality through Fun and Play

Conscientiousness consists of behaviors that go well beyond the minimum role requirements of the organization (Law, Wong, & Chen, 2005). Civic virtue is characterized by behaviors that indicate the employees deep concerns and active interest in the life of the organization (Law et al., 2005). Courtesy has been defined as discretionary behaviors that aim at preventing workrelated conflicts with others (Law et al., 2005). This dimension is a form of helping behavior, but one that works to prevent problems from arising. It also includes the words literal definition of being polite and considerate of others (Organ et al., 2006). Sportsmanship has been defined as a willingness on the part of the employee that signifies the employees tolerance of less-than-ideal organizational circumstances without complaining and blowing problems out of proportion. Organ et al. (2006) further define sportsmanship as an employees ability to roll with the punches even if they do not like or agree with the changes that are occurring within the organization.

Examples of why specific segments work or dont work are described below.

Table 1 - Productivity Game Deployment Matrix The most successful games tap in to core skills and apply them in areas that are outside the regular job
4.1.1 Thought Examples: Where Games Work

Based on our game experiences, described briefly below, games which encourage good organizational behavior (or OCBs), but rely on core skills that most employees have in common, are the most valuable domain for Productivity Games. Since the games rely on core skills, all employees in an organization are able to participate. Additionally, since the behaviour is not closely linked to any individuals job, no ones employment is threatened by the success of another team member. For example, imagine a game that helped sort a complicated list of items. All employees in a given department are familiar with the items, and with how the organization prioritizes its work. This provides a great place for everyone to participate on equal footing. But wrapping the sorting and prioritization work in a game-like interface, all players are given a fair chance to contribute. Games for Learning are a well-established genre of software, and many examples are available in the marketplace for children of all ages. Games in this space work because they focus on the development and growth of the individual. Games are designed to encourage learning, and then test for the learning within the context of play. Players are best rewarded by

Productivity Games: Improving Software Quality through Fun and Play

showing how they have improved themselves, rather than comparing raw completion numbers, which can quickly show disparity between students, but the value of play is not lost.
4.1.2 Thought Examples: Where Games Dont Work

To illustrate where Productivity Games can be unsuccessful, let us provide some example scenarios which might better illustrate possible games. First, imagine a game which encompasses the daily tasks and work of a single employee, Joshua. In the Joshua Game, game play maps directly to the unique skills that Joshua uses to earn his paycheck. Players are given points for doing tasks Joshua would ordinarily do in his work. Some players are able to do most of the tasks Joshua is capable of, and some are limited because they do not have the same unique skills that Joshua has. This presents our first problem: games which exclude players are not in the best interest of the organization. Since Productivity Games require a broad number of players to have an impact, the objective of most games must be to add as many players as possible. Games which rely on actions from the bucket of unique skills inherently limit the breadth of players available to play the game. Back to the example, we find another challenge. At the end of the Do Joshuas Job Game, Joshua hasnt won, will that reflect poorly on his performance in his real job? Will it affect his managers opinion of him? One thing for certain is that Joshua does not feel secure in his job anymore. These two issues provide examples why games focused on unique skill sets are difficult to deploy. Additionally, we see how competitive games focused on in-role behaviors can introduce some awkward situations into the workplace and existing performance review processes.
4.2 Engagement

One indirect consequence of Productivity Games is the increased engagement of employees in the organization. From literature referenced above we know the gamer generation have invested a significant portion of their lives in playing games. And it is interesting to identify some of the attitudes and lessons which this younger generation has taken from playing these games. For example, gamers have learned from games that the cost of failing is very low, and they can always retry, yet from this they expect clear feedback as to what they need to do to change their play in order to succeed later on. From this we can see that the younger generation values a feedback loop and transparency in the consequences. Gamers always expect the game to be fair; otherwise, they will not continue to play. They map these same expectations in a game into their job, expecting the workplace to have transparency and a clear feedback loop. They also expect fairness in how they are treated and in how they should treat others. Finally, games dont demand lengthy reading or studying of manuals in order to play. Most games provide an introductory training mode where the player is given the opportunity to learn what they must know in order to move forward into the game. Similarly, in the workplace, the lengthy corporate memo outlining detailed reasons for organizational priorities carries less impact than is desired. Productivity Games provide an opportunity for an organization to communicate an organization objective or priority in a method that easily meets the needs of this younger generation. In a properly designed game, fairness and transparency are in place. A feedback loop demonstrating success or failure clearly teaches and trains employees how to change their behavior. And finally, instead of a lengthy manual or memo, an employee has the opportunity to
Productivity Games: Improving Software Quality through Fun and Play 10

engage quickly and easily in a training mode which provides the basic information required for the employee to play the game. This isnt to imply that employees are more apt to receive criticism (constructive or otherwise). Rather, because the teaching or coaching is framed in a game, they receive the feedback in a manner they are accustomed to learning from already.
4.3 Language Quality Game

The Windows Language Quality Game has been a successful Productivity Game. It addresses organizational citizenship behaviors by calling on employees within Microsoft to apply their core native language skills to help assess the quality of Windows localization and translation work. The traditional business process uses specific language vendors to perform translation work, and then a secondary vendor to assess the quality. The business challenge has been that, for some languages and locales, finding two independent vendors can be difficult and costly. To address this problem, the Language Quality Game was developed to encourage native speaking populations to do a final qualitative review of the Windows user interface and help identify any remaining language issues. The goal was to ensure a high quality language release and using the diverse population of native language speakers within Microsoft has enabled the pre-release software to be validated in a fun and cost-effective way. The list of Windows languages can be found on Microsoft.comi. Table 2 illustrates the success that the Language Quality Game achieved as run against interim builds of Windows 7. A more detailed description of gameplay can be found online at www.42projects.org, but the goal of the game was to achieve reviews of screenshots and dialogs for translation accuracy and clarity. Native language speakers from across Microsofts diverse, international population were invited to play. The results here demonstrate an immense amount of effort applied to the game.
Game Duration Total Players Total Screens Reviewed (Points Earned) Average Screens per Player Top Player Reviewed Total Defect Reports One Month > 4,600 > 530,000 119 > 9,300 > 6,700

Table 2 Language Quality Game Statistics Success in the game was defined as the number of screen reviews across the 36 languages tested. With the incredible response, most languages had several reviewers provide feedback per screen. Because of the latency in reviewing the feedback, defect reports were not included in players scores. But, for the Windows International Test Team, defect reports were the most valuable output of the game.

Productivity Games: Improving Software Quality through Fun and Play

11

Logistically, the massive amounts of feedback were handled by an international team with tools specially designed to display aggregated feedback. The Moderator role was filled on a per-language basis from the ranks of the international team, and allowed the review of multiple pieces of feedback per screen quickly and easily. Where there was obvious consensus from the game players, a defect report would be created. Reviewed screens lacking consensus were quickly reviewed, but at a lower priority and more quickly, such that the screens with the highest likelihood of fixable defects were handled quickly and efficiently.
4.4 Communicate Hope: A Benefit for Disaster Relief - Giving once helps twice.

Communicate Hope A Benefit for Disaster Relief, is a productivity game that enables Microsoft employees to play by providing ad-hoc and directed feedback on the Microsoft Lync 2010 product. Each player earns points for one of five teams - playing for a disaster relief agency - by accomplishing beta testing related achievements and tasks. At the conclusion of the game, Microsoft donates sponsorship funds to the disaster relief agencies based on the performance of the associated teams. The idea that giving once in terms of trying scenarios and sharing feedback with the team would help twice helping the quality of the product and helping disaster relief efforts. The response to this productivity game-based beta program has been very positive. The number of Microsoft Lync 2010 dogfood program participants has increased from an initial 5,000 in May 2010 to 18,000 in September 2010. During that time, over 10,000 Send Us Feedback comments have been received from participants and key surveys such as Your First Week with Communicator 14 have generated over 4200 responses. Over 47,000 This or That Scenario comparisons have been completed by almost 1600 participants to prioritize which Lync 2010 product scenarios are most important to them. Forrester Research has referenced Communicate Hope in a report on Microsoft Serious Games.26

http://blogs.forrester.com/tj_keitt/10-09-24product_managers_take_note_microsoft_using_serious_games_product_test_and_you_can_too
Productivity Games: Improving Software Quality through Fun and Play 12

26

Figure 3 OC Sonar Communicate Hope Instrumentation

4.5 Elevation of Privilege

Elevation of Privilege (abbreviated "EoP") is a card game developed by Adam Shostack, and is designed to provide a fun and educational introduction to the concepts and practice of Threat Modeling. The basic gameplay is similar to that of many "trick-winning" card games, in which a player leads a card of a particular suit, and other players have to play a card that will match the suit, discard a card of a different suit, or play a card of the declared "trump" suit. The winner of the trick will be the player who plays the highest-value trump card, or if all players played cards from the same suit as the lead player, the player who plays the highest-value card from the led suit is the winner of the trick. The winner of each trick then leads for the next trick until all cards have been used. EoP can be played with the goal of simply accruing tricks, and gaining points for each trick won in this matter - but the purpose of the game is to encourage the players to think of credible threats to an application design, so that these threats can be enumerated, analysed and mitigated. To this end, the suits in the EoP deck are the six elements of the "STRIDE" framework of threats Card decks are available for download http://microsoft.com/security/sdl/eop. The deck contains 74 playing cards in 6 suits: one suit for each of the STRIDE threats (Spoofing, Tampering, Repudiation, Information disclosure, Denial of Service and Elevation of Privilege). Each card has a more specific threat on it.

Productivity Games: Improving Software Quality through Fun and Play

13

Figure 4 - Elevation of Privilege game

CONCLUSION

Humans have played games as far back as we have existed. Animals play games. Games and play are natural activities and can be found almost everywhere. So why dont we play games at work? For some reason, many people dont believe that work can be a game or that games can be used to get real work done. The obstacle may be the competition between game score versus the paycheck and other traditional workplace rewards (promotion, bonus, etc.). Games can be used to attract players in a voluntary situation. The definition of play is that its recreational not required. The game player decides when to stop or start playing. In a traditional sense of work, that uncertainty is unmanageable. However, in todays connected world, the unmanageability can be overcome with scale. Fifty people may decide not to show up today, but another two hundred people did and therefore, we got the work completed. As the demographics of a global and multi-generational workforce continue to change our landscape, we need to think differently about how we engage employees and attract talent to high profile, large scale problems. Software testing is a discipline that has already benefitted from the use of games and game mechanics to attract short term bursts of effort across hundreds, thousands, even millions of players to help assess and improve the quality of software.
6 ACKNOWLEDGEMENTS

Harry Emil, Dan Bean, Joshua Williams, Robert Musson, Alan Page, Adam Shostack, James Rodrigues, Anne Legato, Cari Dick, Lync Test Team, Windows International team.
7 REFERENCES

www.defectprevention.org www.42projects.org www.productivitygames.blogspot.com


Productivity Games: Improving Software Quality through Fun and Play 14

Management Innovation Exchange http://www.managementexchange.com/story/organizational-trust-20-42projects Forbes London Business School Business Strategy Review http://www.managementlab.org/files/site/publications/labnotes/mlab-labnotes-010.pdf CNBC blog http://www.cnbc.com/id/30605923 MSDN Game Theory : http://blogs.msdn.com/b/microsoft_press/archive/2009/07/31/portfolio-selection-and-gametheory-in-defect-prevention.aspx Trusted Advisor Interview http://trustedadvisor.com/trustmatters/747/Trust-Quotes-1Ross-Smith-of-Microsoft TechFlash: Microsoft uses Video Game to help Windows speak like a local Dr. Dobbs Interview http://www.ddj.com/blog/debugblog/archives/2008/02/five_questions_46.html MSDN Tester Spotlight http://msdn.microsoft.com/en-us/testing/bb880947.aspx Video http://download.microsoft.com/download/4/3/b/43ba0ecc-264c-49c4-98b0a84db10bd167/Rosss.wmv Google Testing Blog Using Games to Improve Quality http://googletesting.blogspot.com/2008/06/productivity-games-using-games-to.html Hobson and Holt interview http://www.forimmediaterelease.biz/index.php?/weblog/the_hobson_holtz_report_podcast_400_ november_24_2008/ about 4:10 mins into the show
Microsoft Help and Support: Knowledge Base, List of languages supported in Windows 2000, Windows XP and Windows Server 2003, http://support.microsoft.com/kb/292246
i

Productivity Games: Improving Software Quality through Fun and Play

15

Lets Play Super Rutgers RPG: Interactivity by Proxy in an Online Gaming Culture

Kris Ligman University of Southern California School of Cinematic Arts 25 March, 2011

Kris Ligman Lets Play Super Rutgers RPG - 1 Introduction The literature of game studies has so far attended primarily to examinations of affect and aesthetic, both of which bias a particular assumption about how games are received and in what contexts: that is, by the singular player with her hands on the controller. As should be abundantly clear by the ways in which we refer to gaming culture, however, games are not played nor experienced in this sort of theoretical vacuum. Yet our discourse does not reflect this. This oversight is of course understandably motivated by the conventional ways in which we discuss games as being defined by their interactions with players. However, through the process of interactivity, and especially so in an era where digital record-making is commonplace, a noninteractive record is easily produced which can then be experienced by countless subsequent audiences. This paper will seek to look specifically at one particular result of this coalescence of technologies, the multimedia videogame walkthrough, especially those known as Lets Plays. Much of this paper expounds upon research begun in my 2008 paper for UCLA, Watching the Game: Videogames as a Function of Performance and Spectatorship. I will begin by covering some of the foundational work begun in that paper, in which I discuss motivations for game watching and exhibition, before moving on to how these ideas may be applied to the sharing of web assets through online communities. Background: Of Performance and Spectatorship In Watching the Game, I identified two major taxonomic categories of what I called non-playing videogame audiences, whose in-person presence inarguably shapes play and yet is conventionally left unaccounted for. These two categories I deemed the spectator, who watches competitive play or other tests of skill and behaves similarly to sport spectators, and the

Kris Ligman Lets Play Super Rutgers RPG - 2 passenger, often a sibling or friend whose shotgun perspective in the home media setting allows her to take in a different perspective as well as provide unique feedback to the singleplayer enactor of a games narrative or process.1 In cases of both spectators and passengers, I refer to Sports Fans: The Psychology and Social Impact of Spectators (Wann et al, 2001), where we find the top five reasons for spectatorship of sporting events identified as entertainment value, stress relief, group affiliation, self-esteem benefits, and aesthetic.2 We can trace each of these reasons as existing within videogame watching as well, to varying degrees of importance depending on the setting and the observers relationship to the player. Particularly with respect to spectators, whose observation often takes place in a public setting, watching the gameplay of others doesnt simply excite others to competition or investment, it allows them to share in the actual ritual of spectatorship as a performance of identity. We will see that this performance factors quite strongly into rationales of viewership in the case of video walkthroughs. I think some people see video games as just another form of media, like books or movies. Asking them, Would you like to watch someone play a video game? is like asking them Would you like to watch someone watch a movie? says Maxwell Adams, leading Lets Player for the LP collective Freelance Astronauts.3 Without prompting, he too invokes the sports metaphor, likening game players to athletes, adding, If you like playing the game, then you probably enjoy watching it. [] People love this stuff.4 On the side of the player, there remains some ground to be covered in analyzing why players play for others and in what situations. However, performance is again seen to be quite key, as is the social aspect. I've always liked videogames and I like showing my favorites to others, says Voidburger, 23, one of the few rare female Lets Players; I've played games for friends of mine before I knew what LP was and they were bored. And they didn't wanna play

Kris Ligman Lets Play Super Rutgers RPG - 3 either, because they weren't into videogames. It's nice to play a game for an audience who actually wants to see that.5 While many LPers deny doing the walkthroughs for status or praise, the walkthroughs themselves are something to be taken with a little bit of pride, says Lets Play Archive administrator Baldur Karlsson.6 All Lets Players interviewed placed an emphasis on style, technical standards and accuracy, indicating that presentation is central to the performing players ethos. This emphasis was second only to the frequency with which they mentioned showing and sharing their love of videogames to othersindeed, the motivation that is at the heart of Lets Play. Lets Play: A Genesis (But Not a SEGA One) Videogame walkthroughs have long enjoyed popularity online, largely as textual how-to documents on websites such as GameFAQs.com (founded in 1995 by Jeff Veasey and currently owned by CBS). It is not until the rise of media sharing sites such as Youtube and Photobucket in the mid 2000s that we begin to see multimedia walkthroughs, first in the form of screenshots and, soon after, compressed video clips as well. Lets Plays, as the name implies, emphasizes the sort of collectivity and media share culture popularized by these sites. The term and its practice first appear in the gaming section of the SomethingAwful.com forums (the Awful Forums/SA Forums) in 2005, when various forum members began posting long threads sharing their complete playthroughs of favorite childhood games,7 usually as descriptive text interspersed with screenshots. Beginning with renowned LPer and current Awful Forums moderator slowbeef and his video walkthrough of The Immortal (Electronic Arts, 1990), multimedia walkthroughs which combine video, screenshots, character art, extracted audio and hyperlinks become increasingly common, with Lets Plays like Luisfes

Kris Ligman Lets Play Super Rutgers RPG - 4 walkthrough of Digital Devil Saga (Atlus, 2004), which we will look at in detail later in this paper, becoming the norm. According to Archive administrator Baldur Karlsson, the Lets Play Archive (lparchive.org) was founded in 2007 by Awful Forums user From Earth, since retired, when concerns were raised that memorable walkthrough threads were being lost in the inactive backlogs of Something Awfuls fast-paced message boards. In 2008, Karlsson stepped up into the administration role and in the same year was approached by an editor for the Internet Archive (archive.org), Andrew Armstrong, who offered to host the communitys video walkthroughs on archive.orgs servers.8 This provided Lets Play videos a safe haven from takedown notices by publishers and risk-averse video sharing sites like Vimeo and Youtube, who were starting. As of March 2011, 138 of the Lets Play Archives 433 video walkthroughs, including much of the more popular titles, are hosted on archive.org.9 At present, the Lets Play Archive hosts over 670 videogame walkthroughs for roughly as many titles, many of them obscure, cult and fan favorites. These walkthroughs, still primarily drawn from the Awful Forums (which requires paid memberships to post and should not be taken necessarily as open to participation as, for instance, Youtube), typically attend to fairly stringent technical standards but can vary wildly in terms of format and content, depending on the LPer and the intended audience. The support of the Internet Archive in protecting these documents of gameplay makes its own statement on the perceived value of these videos, which are all too easily interpreted as simple infringement, demystifying or devaluing game properties. On the contrary, as we will see by next looking at the documentation style of several LPs, these walkthroughs represent unique, creative reconfiguration of game texts.

Kris Ligman Lets Play Super Rutgers RPG - 5 Walking Through Walkthroughs: Digital Devil Saga and Max Payne To gain deeper insight into how a Lets Play functions and is received by audiences, let us turn our attention to two specific Lets Plays present on the Archives homepage: Luisfes screenshot and video walkthrough of Atluss 2004 Shin Megami Tensei: Digital Devil Saga and archive administrator Karlssons own video walkthrough of Remedy Entertainments 2001 thirdperson shooter Max Payne. Although quite distinct from one another in terms of format and visuality, both attest to the Lets Play communitys commitment to particular definitions of quality which have helped make their walkthroughs enduring documents. Traveling to Nirvana with Hypertext: Cannibal Hinduism in Digital Devil Saga A relatively obscure but cult favorite for the Playstation2, Digital Devil Saga and its 2005 sequel are offshoot games of Atluss long-running Shin Megami Tensei role-playing game franchise. The games tell a Matrix-esque story of a group of combat AI attempting to defeat their rival tribes and thus ascend to Nirvana, yet instead of ascension they find themselves transported into the real world, which may itself be a second stage of simulation. Characters and locations often allude to elements within Hinduism, as does its theme of ascension or descent through samsara. Picking up on this, Lets Player Luisfe organizes his walkthrough thread around a Hinduist model of spiritual enlightenment, tracking the plots movement across chakra and providing outside links to the mythological concepts the game references. A sampling of one of the chapters in Cannibal Hinduism in Digital Devil Saga illustrates Luisfes format. In Part 6, The Karma Temple, Luisfe shows to his readers protagonist Serphs first visit to Sahasrara, the information nexus at the center of the Junkyard. Opening with screen captures providing an expository shot of the temple, Luisfe then includes a link to Sahasraras Wikipedia entry, describing it as the seventh chakra in Hinduism.10 Further

Kris Ligman Lets Play Super Rutgers RPG - 6 screen captures provide us with stills from a long cutscene in which we meet the leaders of the rival tribes, as well as their particular tattoos which signify their demonic alter-egos. Luisfe subtitles these stills with interpretations of the tattoos to explain their significance according to the game. Cryptic to a fault, Digital Devil Saga benefits from this sort of minimalist commentary, which helps elucidate key concepts but remains for the most part hands-off, allowing the game text to speak for itself. Supplemental video of gameplay and cutscene sequences are linked without audio or textual commentary. Later in the chapter, we come to one of the rare moments where the player is asked to make a verbal response: a choice between asking What is the girls name? and Who is the girl?, both referring to Serah, the young girl who will serve as a catalyst for the plot. This is important, Luisfe notes in his textual commentary beneath the screenshot; We already know the name, so its the other question that matters.11 The second question is highlighted for us. We see no animation indicating the choice made, but the subsequent screenshots in which Angel, the voice commanding the tribes to Nirvana, glitches and speaks in a distorted voiceshows us the implied result of Luisfes decision. In this authored instant, we become privy not just to a choice made by the LPer but his reasons for it, reminding readers that that girl has already been foregrounded for us in previous chapters of the walkthrough. It is a subtle moment, but the way in which Luisfe builds up his procedure and uses commentary only as a supplement coalesces into what Karlsson describes as an added value Lets Play, a walkthrough where the foundational text is presented objectively save for particular procedural enhancements meant to improve comprehension. Coordinated Voyeurism: Editing a Better Max Payne Compared with Luisfes encyclopedic Digital Devil Saga LP, Baldur Karlssons video walkthrough of Remedy Entertainments 2001 third-person shooter Max Payne exhibits a more

Kris Ligman Lets Play Super Rutgers RPG - 7 deliberately authored gameplay experience through several key techniques. Each video chapter is overlaid with Karlssons audio commentary, beginning with a casual address to the unseen audience, as though opening a fireside chat: Hey guys, welcome back to Max Payne.12 The video is a fairly linear documentation of Karlssons playthrough supplemented with gameplay tips, noting enemy AI behaviors and the locations of items as well as occasional summary of the plot. The assertiveness of Karlssons voice in directing the viewers attention is already a departure from Luisfes approach; Karlsson then takes his video one step further by editing his footage creatively, as he describes in interview: I used the developer tools that were still accessible in the game to splice in different camera angles that were as cinematic as I could manage. By cutting out the [HUD], carefully placing the camera and loading/reloading I was able to produce something I feel was more interesting than pure game recording. [] For this game specifically it also allowed me also to cut to an appropriate camera to overhear henchmen conversations, which otherwise would have to be witnessed awkwardly crouched behind a crate.13 In this instance were able to see how a Lets Player is able to use developer tools and other available editing techniques to enhance the unique visuality of the game. Though Karlssons methods break with objective documentation of his own gameplay, he does so in order to enhance the cinematic feel of the games scripted events. These are dramatic moments which can only exist with the creative application of tools by motivated fans seeking to draw out the essence of their play. Toward a General Taxonomy of Lets Play We have seen by looking at the above two walkthroughs that though format and intentions may differ, Lets Plays exhibit an attention to thoroughness and communicating to an

Kris Ligman Lets Play Super Rutgers RPG - 8 audience those aspects of gameplay the LPer deems most valuable. With this in mind, we can identify several key characteristics of the Lets Player and begin to develop an operational taxonomy. The Chronicler Possibly the earliest form of Lets Play was the chronicle, game documentation threads on the Awful Forums which were dedicated to providing a single, comprehensive historiography of a videogame deemed valuable and memorable. As Lets Plays evolved to include more multimedia content such as embedded or hyperlinked video, these walkthroughs began to focus on representing all that there is to see and experience. Luisfes Digital Devil Saga walkthrough, committed to telling the games narrative and supplementing the meaning of its highly referential text, can be seen as a chronicle Lets Play. The Expert/Showman Contrasted with the chronicler, who may or may not be showing off expertise at the same time, the expert or showman Lets Player is seeking to create a document of authority. Skill tips, like those we hear in Baldur Karlssons Max Payne LP, are clearly of this particular sentiment. The expert/showman wants to share not simply a comprehensive play but an idealized version. The presence of the Lets Players personality is also more commonly found in these documents, as in the next type. The Comedian Humor is found frequently in LPs across all categories, but there are also explicitly satirical, sarcastic and oppositional walkthroughs which seek audience attention specifically for their humor. The style of these can vary wildly, from single-player voiceovers reacting to some dated title or recordings of multiplayer plays among friends, such as the often schadenfreude-

Kris Ligman Lets Play Super Rutgers RPG - 9 laden recordings of New Super Mario Bros Wii (Nintendo, 2009) on four-player mode.14 As Karlsson notes, The game was a medium to join in on a shared experience almost. It wasn't about how funny and cool the guys were, but how funny and entertaining the situation and antics were.15 Also of note is that this is the form of Lets Play that has found the greatest success in short form. While most walkthroughs are presented as comprehensive documents, humorous Lets Plays are also enjoyed by viewers in the form of short excerpts and clips. It is also here that we see the greatest potential crossover with machinima, with works like Freemans Mind16 blurring the boundary between what is annotated play and what is pure creative performance. Counter-Histories As a practice of record-making, Lets Play contains some inherent limitations. Chief among them are the ways in which particular records of games crystallize in the minds of players how the game is or should be perceived. Thus, counter-historiographers take it upon themselves to offer alternative documents which allow for the revisiting of gaming artifacts, opening up new interpretations, and adjusting the objective record. One way of examining this practice is to look at instances where two or more Lets Plays exist for a particular title. For instance, the main Lets Play Archive hosts two separate walkthroughs of LucasArtss 1998 Grim Fandango, a 3D point-and-click adventure game notable for lead writer Tim Schafers unique writing style. The first record is a screenshot and textual walkthrough by Luisfe, whose procedural style seen in his Digital Devil Saga LP returns here in the form of minor notes translating Mannys Spanish, but otherwise few interventions.17 When asked in the beginning of his walkthrough why he was covering the same title, second LPer Vexation answers, The original thread was mainly in screen shot format which didn't

Kris Ligman Lets Play Super Rutgers RPG 10 included alot [sic] of the excellent voice work and music that the game features.18 Vexation also cites a personal vested interest in having a video record on hand, fearing that it would be a hassle to have the game play properly on future hardware. Here we see the Lets Player strongly interested in providing his own chronicle for the dual reasons that he feels it is underserved by the existing record and also potentially underserved by technological advancement. Vexations unstated assumption, that his uploaded encoded video will weather these same technological changes much better than the game itself, is also worth noting. Counter-histories are also means by which players can test the limits of their influence on a game engine to record their own preferences or personal stamp upon the game. A good example of this would be the proliferation of video walkthroughs on Youtube which feature the customizable protagonists of Bioware RPGs Dragon Age (2009-2011) and Mass Effect (20072010). Although both franchises have male and female default avatars, the aesthetic choice of a player to adjust their protagonists appearance can in turn become a reflection on the player in a subtle but defining way. It may be especially valuable to further explore how minority representation is emboldened by these personal histories, such as women gamers claiming a sense of empowerment from playing (and documenting themselves playing) the female Commander Shepard never present in Mass Effects official promotional materials. Identity performance is thus far more central to the counter-historiographer Lets Player than the other types and may demand the most rigorous investigation in the future. Furthermore, this is a topic of particular relevance already among certain developers, such as David Gaider, lead writer of the Dragon Age series: We have a lot of fans, many of whom are neither straight nor male, and they deserve no less attention. [] [They] have just as much right to play the kind of game they wish as anyone else.19

Kris Ligman Lets Play Super Rutgers RPG 11 Conclusion The what and why of the Lets Play phenomenon can be located at the heart of Henry Jenkinss 2006 Convergence Culture, which sees the web not simply pulling together media but redefining how that media is engaged. It alters the relationship between existing technologies, industries, markets, genres, and audiences. Convergence alters the logic by which media industries operate and by which media consumers process news and entertainment.20 Lets Plays do not merely create documents of games but develop shareable assets infused with the Lets Players interests and personalities. The amateur digital tools now at the disposal of the vast majority of web users has led to walkthrough content that is raw or refined, esoteric or encyclopedic, gamic as well as non-gamic. While nothing can fully predict how this practice will expand, the crowdsourcing of knowledge found in other fan activities in Convergence Culture would seem to predicate an increased acceptance of games wholly or partially experienced second-hand through documentation. Lets Play as a form of game watching may appeal to spectators and passengers alike, depending on the format and content of the walkthrough. That these walkthroughs have managed to develop in such a diverse fashion over the last few years speaks positively of the way they might still expand and take on additional influences. As noted earlier, the main Lets Play Archive attends to particular technical and content standards which may potentially prove the biggest obstacle to its widespread expansion. Nevertheless, imitators are free to set up their own video walkthrough sites, Youtube channels and personal blogs, handpicking the Lets Play standards they happen to prefer. When asked why the Archives former tagline (Well play it so you dont have to!) is no longer present on the Archives website, Karlsson answered, The motto wasn't actually

Kris Ligman Lets Play Super Rutgers RPG 12 removed consciously really [] But I think also that Let's Plays have moved on from exclusively showing off a game just for people who haven't played it. There's a lot more to LPs now, both in terms of value-adding to a basic playthrough, and in doing something different.21 The name Lets Play itself, a fun invitation to collective media engagement, today does not represent the obviation of play but a welcome message to those already with an interest in gaming who wish to self-educate and explore, making the practice one in which group membership is enhanced and validated. As game platforms diversify and even core titles adopt social characteristics, we will see how the multimedia game walkthrough continues to adapt. Baldur Karlsson, by day a programmer for Crytek UK, waves off attempts to make projections on where games or social media will go. I come from a computing background and I know how often they end up being wrong! he says, noting that while Lets Plays will certainly change depending on how games themselves change, its hard to make specific predictions.22 Nevertheless, he seems optimistic about Lets Plays future, noting he currently receives more volunteer offers than the Archives 18,000 hits-per-day need.23 Although relatively specialized within the larger context of media sharing among online gaming communities, Lets Play is but one example in which the practice of game watching and Web 2.0 websites have interacted to create something far more prevalent and widely recognized than in the past. In the conclusion of our interview, Karlsson turned the investigative lens onto me, asking, How do you think developers currently would feel about LPs were they aware of them? Do you think that LPs could ever be utilized by developers as a tool - either for advertising, or community building, or any other purpose?24

Kris Ligman Lets Play Super Rutgers RPG 13 It serves as both a general question and an overt concern. Should we take a prohibitive stance to a practice which reinvigorates interest in littler-known, often out-of-print titles and puts a creative spin on more popular ones? Do Lets Plays simply demystify and devalue games as products? Are they tantamount to piracy? Or are they potentially some of the best advertising a publisher could ask for?

References
1

Ligman, Kris. Watching the Game: Videogames as a Function of Performance and Spectatorship. dichtungdigital. 2008. Web. <http://www.brown.edu/Research/dichtung-digital/2008/2/Ligman.htm> 2 Wann, Daniel L. et al. Sports Fans: The Psychology and Social Impact of Spectators. New York: Routledge, 2001. 31. 3 Maxwell Adams. Email interview. 15-25 Mar, 2011. 4 Ibid. 5 Voidburger. Email interview. 18-23 Mar, 2011. 6 Karlsson, Baldur. Email interview. 4-22 Mar, 2011. 7 Ibid. 8 Ibid. 9 Lets Play. Internet Archive. Web. 18 Mar, 2011. <http://www.archive.org/details/lets-play> 10 Luisfe. Cannibal Hinduism in Digital Devil Saga. Lets Play Archive. 10 Mar, 2010. Web. 12 Jan, 2011. <http://lparchive.org/Shin-Megami-Tensei-Digital-Devil-Saga/> 11 Ibid. 12 Karlsson, Baldur. Max Payne Part 1 Chapter 2: Live from the Crime Scene. Baldurdash. Web. 20 Mar, 2011. <http://bd.baldurk.org/825#title> 13 Karlsson, Baldur. Email interview. See 3. 14 Maxwell Adams et al. New Super Mario Bros Wii. Freelance Astronauts. 18 May, 2010. Web. 11 Feb, 2011. <http://www.freelanceastronauts.com/games.php?id=52> 15 Ibid. 16 A comedic machinima series by Ross Scott, Freemans Mind began its run in 2007 and continues to be a popular series on Machinima.com. The series is a relatively dressed-down (for machinima) recording documenting Gordon Freemans journey through the locations of the original Half Life (Valve, 1998), supplemented by Freemans inner monologue (provided by Scott). Much of the subversive humor of Freemans Mind, then, is in lending a provocative personality to a traditionally silent protagonist. 17 Luisfe. The Day of the Dead is coming, Lets Play Grim Fandango! Lets Play Archive. 18 Feb, 2007. Web. 3 Mar, 2011. <http://lparchive.org/Grim-Fandango-%28Screenshot%29/> 18 Vexation. Get your tickets ready for Grim Fandango: The Movie. Lets Play Archive. 09 May, 2008. Web. 3 Mar, 2011. <http://lparchive.org/Grim-Fandango-%28Video%29/> 19 Gaider, David. Re: Bioware Neglected Their Main Demographic: The Straight Male Gamer. Bioware Social Network. 19 Mar, 2011. Web. 20 Mar, 2011. <http://social.bioware.com/forum/1/topic/304/index/ 6661775&lf=8> 20 Jenkins, Henry. Convergence Culture: Where Old and New Media Collide. Print. New York: New York University Press, 2006. 16. 21 Karlsson, Baldur. Email interview. See 3. 22 Ibid. 23 Ibid. 24 Ibid.

The Game Behind the Video Game: Business, Regulation, and Society in the Gaming Industry New Brunswick, New Jersey, USA, April 8-9, 2011

For Constitutional Control of New Media: Examining the Judicial Obsession with Causality in Video Software Dealers Association v. Schwarzenegger
Ju Young Lee Penn State University University Park, PA 16802, USA

ABSTRACT On November 2, 2010, the U.S. Supreme Court heard arguments from the State of California, who attempts to adopt a statute that prohibits the sale or rental of deviant and morbid violent video games to minors. In this hearing, the Court appears to recognize possible harmful effects of video games and the necessity for protecting children from that harm. As a result, it might be conjectured that the Court would adjudicate Video Software Dealers Association v. Schwarzenegger (2009) focusing on the legal and practical effectiveness of the state restriction on video games in protecting minors from violent video games rather than on the identification of a compelling government interest. This approach is in line with the decision made by the Court in United States v. Playboy Entertainment Group, Inc. (2000) that struck down Section 505 of the Telecommunication Act of 1996 due to the availability of a less restrictive and allegedly more effective means to shield children from signal bleed. This Playboy Courts approach provides meaningful insights into balancing between controlling negative effects of new media and formulating regulations that do not stifle the development of new forms of media content. The Schwarzenegger case generated that balancing problem by placing too much emphasis on the causality between violent video games and the actual harm for establishing a compelling interest, showing a different approach from the Playboy Court. This study stemmed from the criticism of the courts obsession with the practically unverifiable causality and compared the Playboy decision, as a legal precedent, and the Schwarzenegger decision. Analyzing Schwarzenegger using Playboy lenses will provide a useful framework to evaluate the regulation of new forms of expression. KEYWORDS: videogame, violence, freedom of expression, strict scrutiny, obscenity 1 INTRODUCTION

On November 2, 2010, the U.S. Supreme Court heard arguments from the State of California, which attempted to adopt a statute prohibiting the sale or rental of deviant and morbid, violent video games to minors. In this hearing, the Courts questions revolved around the following three points. First, the Court questioned why violent video games should be treated differently from other forms of media content such as violent films, comic books, fairy tales and rap music, which have First Amendment protection and have never been regulated. Second, it expressed doubt about how the State could determine the deviant violence of video games coupled with the question of the definition of minors. Third, the existence of a compelling government interest was questioned. Specifically, the Court asked questions about whether a category of violent video games, defined by the California statute, is harmful to minors, and whether parents can fully supervise and control what games are being played by their children (Oral Argument before the Supreme Court, 2010).
Lee 1

In short, the Court was concerned with the qualification of violent video games for First Amendment protection, the vagueness of the California law, and the necessity of the governmental intervention in video game sale or rental to minors. These concerns seem to be merely an extension of the questions that were considered in Video Software Dealers Association (VSDA) v. Schwarzenegger (2009) and repeatedly contested in other Circuit and district courts that have struck down the similar laws restricting minors access to violent video games.1 However, the Court, or more precisely, some of the Justices, seemed to recognize possible harmful effects of video games and the necessity for protecting children from that harm. As a result, it might be conjectured that the Court would adjudicate the Schwarzenegger case focusing on the legal and practical effectiveness of the state restriction on video games in protecting minors from the deviant and morbid violence of certain video games, rather than on the identification of a compelling government interest. This approach is in line with the decision made by the Court in United States v. Playboy Entertainment Group, Inc. (2000). In that case, the Court struck down Section 505 of the Telecommunications Act of 1996, due to the availability of a less restrictive and allegedly more effective means to shield children from signal bleed, the partial reception of video images and/or audio sounds on a scrambled [sexually explicit adult programming] channel (Playboy Entertainment Group, Inc. v. United States, 1998, p. 706). The Playboy Court also questioned whether a compelling government interest is achieved by imposing the blocking, scrambling or time-channeling requirements on cable operators, which provide sexually-oriented programming. Yet, the Courts question does not necessarily mean that it doubts the harmful effect that indecent content might have on minors. Its concern was in whether the signal bleed problem, which must be addressed by government regulation, actually exists.2 In relation to the debate on the protection of minors from arguably harmful violent video games, the implications of the decision in Playboy are twofold. First, in deciding to shield children from certain speech protected by the First Amendment, the Court considered parents concern about their childrens access to the highly offensive material. Second, when there are alternative ways to achieve a compelling interest, the government must prove that the alternatives are less effective and not less restrictive than the regulation at issue. The Playboy Courts approach provides meaningful insights into balancing between controlling the negative effects of new media and formulating regulations that do not stifle the development of new forms of media. Video Software Dealers Association (VSDA) v. Schwarzenegger (2009) is a case involving the balancing problem. In Schwarzenegger, the U.S. Court of Appeals for the 9th Circuit affirmed that, under the standard of strict scrutiny, the California statute is unconstitutional because the state government failed to demonstrate the causal relationship between violent video game playing and psychological or neurological harm to minors, and because the government failed to prove that there exists no less restrictive alternative to further the purported government interest (VSDA v. Schwarzenegger, 2009, p. 967). Although the Court also considered the overbreadth of the Act, the key reason for its holding appears to consist in whether the State successfully established a compelling interest by providing sufficient evidence that violent video games cause actual harm to minors. With regard to this decision, some important questions can be asked. Is the video game in which the minor commits violent acts of maiming, killing, setting people on fire not harmful to minors? (Oral Argument before the Supreme Court, 2010) Is the actual harm of a violent video game testable or verifiable? Under strict scrutiny, is the least restrictive means standard only an element of the narrowly tailored means standard? (Is there a
1

There are four circuit court cases on video game laws including the Schwarzenegger case: American Amusement Machine Association v. Kendrick (2001), Interactive Digital Software Association v. St. Louis County (2003), and Entertainment Software Association v. Blagojevich (2006). Additionally, there are several district court cases that have struck down similar video game restrictions: Video Software Dealers Association v. Maleng (2004), Entertainment Software Association v. Foti (2006), Entertainment Software Association v. Granholm (2006), Entertainment Merchants Association v. Henry (2006), and Entertainment Software Association v. Swanson (2006). 2 The existence of the signal bleed problem was not a critical factor in the decision.
Lee 2

possibility that the least restrictive means standard could be an independent factor determining the constitutionality of the speech restriction?) This paper addresses these questions by comparing the Playboy decision, as a legal precedent, and the Schwarzenegger case. The Court mentioned United States v. Playboy Entertainment Group, Inc. only as a precedent that applied the least restrictive means standard to strike down a government restriction on speech. Nevertheless, the implications of the Playboy decision for the evaluation of the constitutionality of video game regulations seem more than that. Analyzing Schwarzenegger using Playboy lenses would provide a useful framework to evaluate the regulation of new forms of expression. 2 CONTENT-BASED RESTRICTION AND STRICT SCRUTINY

In First Amendment jurisprudence, restricting speech, which has First Amendment protection based on its content, is presumptively unconstitutional, and that content-based restriction is subject to strict scrutiny. Therefore, in order for a statute restricting protected speech to be valid and constitutional, the government must demonstrate that the statute is necessary to serve a compelling state interest and is narrowly drawn to achieve that end (Arkansas Writers Project, Inc. v. Ragland, 1987, cited in Franklin, Anderson, & Lidsky, 2005, p. 31) However, according to Franklin, Anderson, & Lidsky (2005), the least restrictive means standard is not always applied by the Court as the third requirement. In many cases, whether the government regulation is the least restrictive means has been examined to determine the narrowly tailoredness of the regulation. In this section, the facts and decision of Video Software Dealers Association v. Schwarzenegger (2009) and United States v. Playboy Entertainment Group, Inc. (2000) will be reviewed for a comparative analysis of the two cases. 2.1 Video Software Dealers Association v. Schwarzenegger (2009) In October, 2005, the state of California enacted a law (Civil Code 1746-1746.5), which states that, [a] person may not sell or rent a video game that has been labeled as a violent video game to a minor (California Civil Code 1746.1(a)). The Act defines a violent video game as follows: (d)(1) "Violent video game" means a video game in which the range of options available to a player includes killing, maiming, dismembering, or sexually assaulting an image of a human being, if those acts are depicted in the game in a manner that does either of the following: (A) Comes within all of the following descriptions: (i) A reasonable person, considering the game as a whole, would find appeals to a deviant or morbid interest of minors. (ii) It is patently offensive to prevailing standards in the community as to what is suitable for minors. (iii) It causes the game, as a whole, to lack serious literary, artistic, political, or scientific value for minors. The VSDA argued that the Act is unconstitutional on the grounds that video games are speech protected by the First Amendment, and that the definition of violent video game provided in the Act was unconstitutionally vague, making its enforcement impossible. They also asserted that the Act was not narrowly tailored because there is a voluntary rating system, implemented by the Entertainment Software Rating Board (ESRB), which provides information about video game content and therefore helps parents control what games their children can purchase and play (VSDA v. Schwarzenegger, 2009). On the other hand, the State argued that the Act is allowed by the First Amendment based on the variable obscenity or obscenity as to minors standard established in Ginsberg v. New York (1968). Specifically, the State asserted that there is a compelling government interest in [preventing] harm to children and [enabling] parents to guide their childrens upbringing, and that the Act is the least restrictive means to further the States compelling interest because the ESRB self-regulating rating system a voluntary system without the force of law or civil penalty and the technological system for parental control may not adequately protect minors from unquestionably violent video games (VSDA v. Schwarzenegger, 2009; Petitioners Brief, 2009).
Lee 3

Despite the States arguments, the 9th Circuit held that the Act violates rights guaranteed by the First and Fourteenth Amendments. The court asserted that the California statute is a content-based restriction and thus subject to strict scrutiny and not the variable obscenity standard from Ginsberg. Applying strict scrutiny, the court found that the State did not demonstrate a compelling interest in preventing psychological or neurological harm of violent video games to minors, and that even if the State established a compelling interest, the Act was not narrowly tailored because there exist less restrictive means to promote the purported interest. The labeling requirement was also held to be unconstitutional. The court held that the mandatory labeling is impermissibly compelled speech under the First Amendment because the label would provide the States controversial opinion rather than purely factual information. However, the court did not uphold VSDAs argument that the definition of violent video game is unconstitutionally vague (VSDA v. Schwarzenegger, 2009, p. 953). 2.2 United States v. Playboy Entertainment Group, Inc. (2000) In 1996, Congress passed the Communications Decency Act (CDA), which is Title V of the Telecommunications Act of 1996. Section 505 of the Act required cable television operators providing channels primarily dedicated to sexually-oriented programming either to fully scramble or otherwise fully block those channels or to broadcast them during the "safe-harbor" hours of 10 p.m. to 6 a.m. when children are unlikely to be viewing (Telecommunications Act of 1996 505, 47 U.S.C. 561). Although scrambling was already being used by cable operators to allow only subscribers to watch the premium channels, either or both audio and visual portions of the scrambled programs might be heard or seen. Section 505 was enacted to protect children from that signal bleed problem resulting from imperfect scrambling (United States v. Playboy Entertainment Group, Inc., 2000, P. 806). Playboy Entertainment Group, Inc. challenged the constitutionality of Section 505 by arguing that the Act is an unnecessarily restrictive content-based regulation violating the First Amendment. The main complaint about Section 505 is related to the fact that most cable operators providing adult-oriented programming channels had no practical choice but to adopt the time-channeling option so as to avoid the penalties imposed if any audio or video signal bleed occur during the regulated sixteen hours except the safe-harbor hours. The Court recognized that the extensive adoption of time channeling, in effect, meant the removal of sexually explicit adult programming from the time slot outside the safeharbor period, and that the elimination of adult programming infringes on adults constitutional right to view it, as well as speakers right to adult speech (United States v. Playboy Entertainment Group, Inc., 2000, p. 807, 811). It was assured that the statute must be narrowly tailored to respect adults' viewing freedom in order to restrict no more speech than necessary, and that the State must choose, if any, an alternative means that would further the compelling interest in a less restrictive but at least as effective way. The Court found that Section 504, which requires a cable operator to fully scramble or block any channel a subscriber does not want to receive upon the subscribers request (Telecommunications Act of 1996 504, 47 U.S.C. 560), would provide as much protection against unwanted programming as would Section 505 and be less restrictive on Playboy's First Amendment rights because it was a contentneutral regulation (United States v. Playboy Entertainment Group, Inc., 2000, p. 810-813). Based on the reasoning above, the Court concluded that Section 505, a content-based restriction on indecent speech protected by the First Amendment, was unconstitutional because the Government failed to demonstrate that Section 505 was the least restrictive means to achieve the compelling government interest, and was narrowly tailored. 3 PROTECTION OF MINORS AND FREEDOM OF EXPRESSION

The two cases reviewed are representative cases holding the unconstitutionality of government regulation of a new form of media content for the purpose of protecting minors from allegedly harmful speech. These cases are similar in that they address the constitutionality of the regulations intended to protect minors (or children) from the possible harm of controversial media content such as indecent
Lee 4

programming and violent video games. Access to the controversial content at issue in these cases is very limited. That is, video games should be purchased or rented for playing, and the adult-oriented programming is scrambled and offered on a subscription basis. Accordingly, the harm of these materials to minors might be prevented by parental controls of minors access to them. Nevertheless, the government, under the assumption that parents might not fully supervise and control what their children play and watch, enacted the law that directly restricts the speech at issue to facilitate parental controls. The validity of the governmental restriction on speech depends on whether the speech at issue has First Amendment protection and whether the restriction is content-based. If the restriction is a contentbased regulation of protected speech, it is presumptively unconstitutional and must pass strict scrutiny to be valid. Schwarzenegger and Playboy will be analyzed based on this First Amendment principle. However, in applying strict scrutiny, the least restrictive means standard will be regarded as the third prong independent of the narrowly tailored means standard. The unique implication of the least restrictive means standard will be shown in the following analysis. 3.1 Qualification of Violent Video Games for First Amendment Protection In VSDA v. Schwarzenegger (2009), the State pointed out the obscene level of violence (Oral Argument before the Supreme Court, 2010) of certain video games and argued that violent video games were not protected speech, which could be constitutionally restricted by the California statute. However, the court declined the States application of the Ginsberg rationale the variable obscenity or obscenity as to minors standard to violent video games by quoting the Ginsberg Courts holding that shows the Courts intention to limit the application of the variable obscenity standard: We have no occasion in this case to consider the impact of the guarantees of freedom of expression upon the totality of the relationship of the minor and the State. It is enough for the purposes of this case that we inquire whether it was constitutionally impermissible for New York . . . to accord minors under 17 a more restricted right than that assured to adults to judge and determine for themselves what sex material they may read or see. (VSDA v. Schwarzenegger, 2009, p. 960) In addition, the court also cautioned against expanding the boundaries of the legal concept of obscenity under the First Amendment based on the opinion in Miller v. California (1973) that expressly limited the permissible scope of regulation to sexually explicit material (VSDA v. Schwarzenegger, 2009, p. 960). In American Amusement Machine Association v. Kendrick (2001), which involved a restriction of violent and sexual video games, the Seventh Circuit distinguished violence from obscenity. Although the State constructed the definition of violent video games by incorporating the three prongs of the Miller standard, video game violence should not be interpreted or treated as obscenity in consideration of the Courts original intention. Most video games, even extremely violent video games, are based on plot lines and could have some artistic values. Moreover, there is no historical tradition that violent speech was regulated (Oral Argument before the Supreme Court, 2010). Therefore, violent video games should be assumed as protected speech, and the California video game law is subject to strict scrutiny. 3.2 Compelling Government Interest When imposing a restriction on protected speech based on content, the government must first demonstrate that there exists a compelling government interest in regulating the speech. In Schwarzenegger, the court questioned the existence of the States interest that justified the protection of minors from the alleged harm of violent video games. Based upon the decision of Turner Broadcasting Sys., Inc. v. FCC (1994), the court required the government to demonstrate that the recited harms are real, not merely conjectural (Turner Broadcasting Sys., Inc. v. FCC, 1994, p. 664, cited in VSDA v. Schwarzenegger, 2009, p. 962 ). Considering a number of documents legal analyses, general background papers, position papers, etc. including studies on the negative effect of violent video

Lee

games (i.e., Anderson, 2004; Gentile, 2004; Funk et al., 2004), the 9th Circuit found that those studies did not provide sufficient evidence of a causal link between playing violent video games and psychological or neurological harm to minors, as other courts found (e.g., American Amusement Machine Association v. Kendrick, 2001; Interactive Digital Software Associationn v. St. Louis County, 2003; Video Software Dealers Association v. Maleng, 2004; Entertainment Software Association v. Granholm, 2006; Entertainment Software Association v. Blagojevich, 2006) (VSDA v. Schwarzenegger, 2009, p. 963-964). However, the courts rejection of the existence of the compelling government interest appears to overlook an important interest that might justify the States regulation of the sale or rental of violent video games. In United States v. Playboy Entertainment Group, Inc. (2000), the Court assumed that: many adults themselves would find the material highly offensive, and [that] when the material comes unwanted into homes where children might see or hear it against parental wishes or consent, there are legitimate reasons for regulating it. (p. 811) This indicates that parents concern about their childrens access to harmful speech may constitute a compelling government interest. In protecting children from the signal bleed of sexually-oriented programming channels, the proof of the causal relationship between sexually explicit material and actual harm to minors was not required. In Ginsberg, the Court considered the parents concern, as well as the states independent interest in the wellbeing of its youth, in justifying the application of the lowest level of scrutiny (i.e., rational basis standard) to the prohibition of the sale of sexually-explicit material to minors. The Court recognized that constitutional interpretation has consistently recognized that the parents claim to authority in their own household to direct the rearing of their children is basic in the structure of our society. (Ginsberg v. New York, 1968, p. 639-640, cited in VSDA v. Schwarzenegger, 2009, p. 959) While it has been recognized that sexually-oriented indecent speech is harmful to minors, there is no consensus on the harm of violent speech. However, it is practically impossible to test or verify the causal effect of both sexual and violent speech on minors because longitudinal studies of that effect would involve exposing minors to the allegedly harmful content. Therefore, instead of requiring unobtainable proof of actual harm of violent video games, the Court should consider the parents concern about their childrens access to highly offensive violent video games. Additionally, the characteristics of video games as a new form of media content should not be overlooked. In the Supreme Court hearing of the oral argument for VSDA v. Schwarzenegger, some Justices expressed serious concern over violent video games that enable the minor to commit violent acts, such as killing, maiming, and sexually assaulting virtual human beings, setting them on fire, and urinating on them (Oral Argument before the Supreme Court, 2010). This simulation feature of video games has been regarded as a unique characteristic that makes video games an effective tool to educate or develop a certain skill (e.g., Aldrich, 2009; Smith, 2009). That simulation feature and the interactivity of video games could support the conclusion of many studies that violent video games may have a larger impact on minors behavior and cognition than violent movies, books, and songs. That conclusion also supports the States argument that sound policymaking often requires legislators to forecast future events and to anticipate the likely impact of these events based on deductions and inferences for which complete empirical support may be unavailable. (Turner Broadcasting System, Inc. v. FCC, 1994, at 665, cited in Petitioners Brief, at 50) In conclusion, parents concern about the possible negative impact of violent video games needs to be viewed as an important factor that determines whether there is a compelling government interest in restricting the minors access to violent video games. The dissenting opinion in Playboy asserted that the [g]overnment has a compelling interest in helping parents by preventing minors from accessing sexually explicit materials [which are assumed to be harmful to minors] in the absence of parental supervision. (United States v. Playboy Entertainment Group, Inc., 2000, p. 8433) Accordingly, the [g]overnment disclaims any interest in preventing children from seeing or hearing [them] with the consent of their
3

Justice Breyer, with whom The Chief Justice, Justice O'Connor, and Justice Scalia join, dissenting.

Lee

parents. (United States v. Playboy Entertainment Group, Inc., 2000, p. 811) By the same token, the California state could argue that there is a compelling interest in empowering parents to control their childrens access to allegedly harmful violent video games by banning the sale or rental of violent video games. 3.3 Narrowly Tailored Mean In both Playboy and Schwarzenegger, the government restrictions on speech were found to be not narrowly tailored because less restrictive alternatives to achieve the government end existed. However, whether the government regulation at issue meets the narrowly tailored means standard can be examined from a different perspective. In Video Software Dealers Association v. Maleng (2004), the court ruled that the statute, prohibiting the distribution to minors of video games involving violence against law enforcement personnel, was both over-inclusive and under-inclusive because the set of [video] games covered by the statute did not reflect the harms the legislature sought to alleviate; the statute was therefore not narrowly tailored. (Video Software Dealers Association v. Schwarzenegger, 2005, p. 1044). This case indicates how the narrowly tailored means standard can be distinguished from the least restrictive means standard. A narrowly tailored means may mean a measure that effectively promotes the government goal by restricting only targeted speech but does not affect other legitimate speech or rights. Taking this approach, in relation to the decision of Playboy and Schwarzenegger, three questions can be asked: 1) can the statute achieve the governments child-protecting objective?; 2) is the definition of the targeted speech not vague?; and 3) does the statute infringe adults right to the targeted speech? In Playboy, Section 505 provides a clear definition of the targeted speech, signal bleed, which means the partial reception of video images and/or audio sounds on a scrambled [sexually explicit adult programming] channel (Playboy Entertainment Group, Inc. v. United States, 1998, p. 706). However, the Court doubted Section 505s efficacy in achieving the compelling interest of protecting children from signal bleed under the premise that most cable operators complied with Section 505 by time-channeling their programs. The Court found that the time channeling requirement may not prevent minors from being exposed to inadvertent signal bleeding of adult programming channels in the absence of parental supervision. Furthermore, time-channeling adult-oriented programs may result in eliminating those programs from the time period between 6 a.m. and 10 p.m. and thereby infringe on adults constitutional right to view them, as well as Playboys First Amendment right to transmit adult speech (United States v. Playboy Entertainment Group, Inc., 2000, p. 807, 811). In sum, Section 505 is not narrowly tailored for furthering the compelling interest. The Schwarzenegger decision can be analyzed using the same frame. Although VSDA continuously argue that the definition of violent video game covered by the Act is unconstitutionally vague, the court did not uphold their argument. The court held that the severability clause in Civil Code 1746.54saved the Act despite the unconstitutionally broad definition of violent video game in Section 1746(d)(1)(B)5 (VSDA v. Schwarzenegger, 2009, p. 953-956). The court did not provide a clear opinion on the effectiveness of the Act in achieving the States interest in protecting minors from psychological or neurological harm, but it appears to assume that the ban on the sale or rental of violent video games to minors and the labeling requirement would keep violent video games from being legally purchased by anyone under 18 years of age (VSDA v. Schwarzenegger, 2009, p. 966). This statute, in principle, would not likely affect the adult right to play those video games. However, in practice, because retailers might not like to sell labeled video games, this retailers tendency could reduce the number of video games that are available to adults, indirectly infringing both the adults right to play video games and the video
Civil Code 1746.5 states that [t]he provisions of this title are severable. If any provision of this title or its application is held to be invalid, that invalidity shall not affect other provisions or applications that can be given effect without the invalid provision or application. 5 Section 1746(d)(1)(B): Enables the player to virtually inflict serious injury upon images of human beings or characters with substantially human characteristics in a manner which is especially heinous, cruel, or depraved in that it involves torture or serious physical abuse to the victim.
Lee 7
4

game producers right to communicate with adult video gamers through video games. The court also pointed out that the Act does not speak about an exception for sales to minors accompanied by a parent, restricting the minors access to video games that are allowed by parents (VSDA v. Schwarzenegger, 2009, n.4). With regard to the labeling requirement, the court hinted that the labeling could make legitimate video games prevented from being sold to minors by holding that the label provides the States controversial opinion rather than purely factual information. (VSDA v. Schwarzenegger, 2009, p. 953) Despite the problems examined above, the court, unlike the Playboy Court, did not specify those problems in its opinion by evaluating whether the Act was narrowly tailored mainly in terms of the availability of less restrictive means to promote the compelling interest. This leaves many questions that should be addressed in the Supreme Court. 3.4 Least Restrictive Means There appears to be no disagreement on the principle that the existence of less restrictive alternatives presumes the over-inclusiveness of the measure at issue and automatically constitutes a violation of the narrowly tailored means standard. Nevertheless, this third prong needs to be discussed separately because it is very difficult to determine whether or not a means is the least restrictive and, at the same time, most effective means. Also, this standard is a safety valve that prevents the unduly restrictive regulation of new forms of media, such as the Internet and video games. In United States v. Playboy Entertainment Group, Inc. (2000), it seems evident and clear that there was a less restrictive alternative to Section 505 requiring full scrambling, full blocking, or time channeling because the statute included Section 504 providing free targeted blocking, (or voluntary blocking) which is a content neutral regulation. The Court held that Section 504 was a less restrictive and more effective means to promote the compelling government interest in that it facilitated parental controls without affecting the First Amendment interests. (p. 815) However, the relative efficacy of Section 504 compared to Section 505 especially the time channeling requirement was controversial. The Court recognized that voluntary blocking might not be effective because it required parents to take action with adequate information. If parents fail to ask cable operators to block the programming, the compelling interest in protecting children from signal bleed would not be served. In determining whether the less restrictive means is feasible and effective, the key point is that it is the governments obligation to prove that the alternative will be ineffective to achieve its goals. (p. 816) Applying those principles to VSDA v. Schwarzenegger (2009), the government must prove that the alternative means, such as the improved ESRB rating system, enhanced educational campaigns, and parental controls, are less restrictive and more effective alternatives to a ban on the sale or rental of violent video games (p. 967). Also, the Court asserted [i]f a less restrictive alternative would serve the Governments purpose, the legislature must use that alternative. (United States v. Playboy Entertainment Group, Inc., 2000, p. 813) The importance of the least restrictive means standard consists in finding less restrictive and more effective means to control the problem and, at the same time, not to stifle the development of new media and new forms of content. Contrary to Playboy, the Schwarzenegger court did not discuss the comparative restrictiveness and effectiveness of the alternative means compared to the proscription of violent video game sales and rentals. This shows that the court struggled to protect First Amendment principles rather than to find a better way to manage a new form of media and to cope with the problem related to the new media. In deciding the video game cases, including Schwarzenegger, the least restrictive means standard would play an important role for establishing a system where freedom of expression and government control of the side effects of the new media can harmonize. 4 CONCLUSION Prior court cases dealing with government regulation of violent video games including VSDA v.

Lee

Schwarzenegger (2009) demonstrate that the court attempted not to expand the boundaries of unprotected speech (i.e., obscene speech) by requiring the government to provide a clear link between video games and the actual psychological and neurological harm to minors. Although applying the Ginsberg rationale to justify the government regulation of violent video games is inappropriate, the underlying purpose of regulating sexually explicit material deserves to be considered in deciding Schwarzenegger. As a new form of speech, video games are challenging many existing rules and necessitating new rules in diverse areas. The First Amendment is one of the rules that are challenged by video games. However, the core value of the First Amendment should not be compromised by the institutional control over the new media. Instead, it is important to find the way to harmonize free speech rights and the compelling government interests such as protecting minors from video game violence. Even though the court strove to preserve the existing rule, Schwarzenegger can be regarded as an attempt to establish a new system and rule. United States v. Playboy Entertainment Group, Inc. (2000) provides useful insights into such rule-making. The Playboy decision suggests that the Supreme Court should encourage the government to formulate more effective regulation by reviewing the Schwarzenegger case mainly in terms of the narrowly tailored means and the least restrictive means standards, as well as the compelling government interest standard. The courts obsession with the causal relationship between violent video games and the alleged harm to minors may frustrate societys effort to develop efficient tools to manage new media and technologies. 5 ACKNOWLEDGEMENTS

I would like to express my special thanks to Dr. Richard Taylor, Palmer Chair of Telecommunications Studies and Law and Co-Director of the Institute for Information Policy at Penn State University, for providing invaluable guidance in writing this paper. 6 REFERENCES Aldrich, C. A. (2009). The Complete Guide to Simulations and Serious Games, San Francisco, CA: Wiley (Pfeiffer). Anderson, C. A. (2004). An Update on the Effects of Playing Violent Video Games, Journal of Adolescence, 27, 113-122. Franklin, M. A., Anderson, D. A., & Lidsky L. B. (2005). Mass Media Law: Cases and Materials (7th ed.). New York, NY: Foundation Press. Funk, J. B., Baldacci, H. B., Pasold, T., & Baumgardner, J. (2004). Violence Exposure in RealLlife, Video Games, Television, Movies, and the Internet: Is There Desensitization?. Journal of Adolescence, 27, 23-39. Gentile, D. A., Lynchb, P. J., Linderc, J. R., & Walsh, D. A. (2004) The effects of violent video game habits on adolescent hostility, aggressive behaviors, and school performance, Journal of Adolescence, 27, 5-22. Murray, J. P. (2008). Media Violence: The Effect Are Both Real and Strong. American Behavioral Scientist, 51(8), 1212-1230. Saunders, K. W. (1996). Violence As Obscenity: Limiting the Medias First Amendment Protection. Durham, London: Duke University Press. Smith, R. D. (2009). Military Simulation & Serious Games. Orlando, FL: Modelbenders Press. Oral Argument before the Supreme Court: Schwarzenegger v. Entertainment Merchants Association. (Nov. 2, 2010). Retrieved from http://www.supremecourt.gov/oral_arguments/argument_audio_detail.aspx?argument=081448 Brief for Petitioners Arnold Schwarzenegger, Governor of the State of California and Edmund G. Brown, Jr., Attorney General of the State of California. Retrieved from http://www.scotusblog.com/case-files/cases/eanf/
Lee 9

Kuchera, B. (n.d.). 72% of adults support gaming laws? Well, kinda... Retrieved September 17, 2010, from http://arstechnica.com/gaming/news/2010/09/72-of-adults-support-gaming-lawswell-kinda.ars American Amusement Machine Association v. Kendrick, 244 F.3d 572 (7th Cir. 2001). Entertainment Software Association v. Blagojevich, 404 F.Supp.2d 1051 (N.D. Ill. 2005). Entertainment Software Association v. Blagojevich, 469 F.3d 641 (7th Cir. 2006). Entertainment Software Association v. Foti, 451 F. Supp. 2d 823 (M.D. La. 2006). Entertainment Software Association v. Granholm, 426 F Supp 2d 646 (E.D. Mich. 2006). Entertainment Software Association v. Swanson, 519 F.3d 768 (D. Minn. 2006). Entertainment Merchants Association v. Henry, 2006 WL 2927884 (W.D. Okla. 2006). Ginsberg v. New York, 390 U.S. 629 (1968). Interactive Digital Software Associationn v. St. Louis County, 329 F 3d 954 (8th Cir 2003). Playboy Entertainment Group, Inc. v. United States, 30 F. Supp. 2d 702 (D. Del.1998). Reno v. ACLU, 521 U.S. 844 (1997). Turner Broadcasting System, Inc. v. FCC, 512 U.S. 622 (1994). United States v. Playboy Entertainment Group, Inc., 529 U.S. 803 (2000). Video Software Dealers Association v. Maleng, 325 F. Supp. 2d 1180 (W.D. Wash. 2004). Video Software Dealers Association v. Schwarzenegger, 401 F. Supp. 2d 1034 (N.D. Cal. 2005). Video Software Dealers Association v. Schwarzenegger, 556 F.3d 950 (9th Cir. 2009). to ortizj@rutgers.edu by March 25th 2011.

Lee

10

The Game Behind the Video Game: Business, Regulation, and Society in the Gaming Industry New Brunswick, New Jersey, USA, April 8-9, 2011

The Private Regulation Of Virtual Worlds Versus Real World laws: European And American Perspectives
Sevan J. Antreasyan University of Geneva 12 Avenue Edmond-Vaucher, 1203 Geneva, Switzerland

ABSTRACT As the popularity of virtual worlds increases all over the world, this paper proposes to evaluate the impact of End-User License Agreements (EULAs) of the most popular virtual worlds on their users, whether they live in the United States or in Europe. It may be recalled that, to enter a virtual world, every user has to accept its EULA. These agreements which are binding as soon as they are accepted contain a variety of rules that apply to users as if they were residents in a new jurisdiction. These circumstances raise various sets of questions that will be addressed in this paper. What is the nature of virtual world EULAs? How do these agreements interact with real world law? Does one set of regulations prevail over the other? To what extent does American companies private regulation apply to European users? Can European Union (E.U.) regulation limit the enforceability of EULAs for its residents? What would be the consequences of such limitations for virtual world providers? By answering these questions, the purpose of this paper is to draw attention to the conflicts between two different sets of regulation i.e. the private regulation of virtual worlds and the real world law and their effects on users rights. These questions are interesting both on a scholarly level (the debate concerning private regulations) and on a practical level (concerning residents rights limitations). Moreover, the validity analysis of EULAs under E.U. regulations could prove important to virtual world providers, given the growth of the virtual world market in Europe. The consideration of E.U. regulations in the drafting of EULAs would most certainly prove to be beneficial for these entities in the future. KEYWORDS: private regulation, real world law, state regulation, EULA, Europe, consumer protection, unconscionability, conflict-of-law, jurisdiction, intellectual property 1 INTRODUCTION

At a time when virtual worlds are expanding their user base all around the world, it is interesting to observe that most of them are operated by companies in the U.S. (Jankowich, 2006). This fact certainly influences how contracts regulating these virtual worlds are drafted and since they do not have any real world boundaries raises questions concerning the enforceability of these rules in other countries, and more specifically in the E.U. Essentially, the game behind EULAs involves two players. The provider of the virtual world (player 1), on the one hand, drafts the EULA clearly in favor of its own interests, and as aggressively as U.S. standards allow. The user of the virtual world (player 2), on the other hand, does not have much choice but to accept the non-negotiated contract if wanting to enter the virtual world. If only the rules of the game were applied, then users would always be on the losing side, due to the inequitable provisions in these contracts. Fortunately, for users of virtual worlds, real world law acts as a referee that has the power to balance some of these inequities. The goal of this paper is to evaluate the role of this referee and the scope of its powers. After a brief presentation of the governance of virtual worlds, underlining the importance of EULAs (part two), conflicts between the private regulation of virtual worlds and real world laws will be analyzed. In this

Antreasyan

respect, the third part of this paper will briefly present conflicts with U.S. regulation, while the fourth section will focus on E.U. regulation. 2 THE GOVERNANCE OF VIRTUAL WORLDS

Virtual worlds are essentially governed by two kinds of regulation: private (2.1) and public (2.2) (see e.g. Jankowich, 2006; Quinn, 2010; Sites et al., 2010). Two additional sources of regulation should also be mentioned: (computer) code and community rules (see e.g. Risch, 2009). However, neither will be considered in this paper. The former which allows the provider to directly control the behavior of users and/or objects in the virtual world will not be considered because the legal issues it raises are not contractual (See generally Lessig, 1999). The latter (community rules) will not be considered because they do not directly concern the relationship between the virtual world provider and the user, but rather the relationship between users. 2.1 Private regulation The first level of regulation is contractual (Duranske, 2008), i.e. private. In order to enter a virtual world, users must agree to the EULA, which the providers of virtual worlds have previously drafted. Users are bound by this non-negotiated agreement once they accept it (for more information on the acceptance of EULAs, see e.g. Sites et al., 2011). EULAs contain all sorts of rules that govern the vertical relationship between the virtual world provider and the user (B2U). The majority of these EULAs also tend to include behavior provisions, which can affect the relationship between users (U2U). Even though this aspect is interesting, it exceeds the scope of this paper and will thus not be considered; the discussion will rather focus solely on the vertical relationship. Although the majority of authors argue that EULAs are not appropriate for comprehensively and exclusively regulating virtual worlds (see e.g. Fairfield, 2008; Lastowka, 2010), these contracts would be used by courts throughout the world as the primary source of law when they deal with cases related to virtual worlds. The freedom of contract doctrine anchored in the Fifth and Fourteenth Amendments of the U.S. Constitution and in all E.U. Member States legal systems (see e.g. Art. 1134.1 of the French civil code and Art. 1255 of the Spanish civil code) provides that agreements legally formed shall be held as the law of the parties that made them. Providers view EULAs as a great tool to allocate rights e.g. intellectual property rights (IPR) and virtual property rights and to limit as much as possible their responsibility toward users, as they can be tailored to their specific needs (Sites et al., 2010). To date, U.S. courts have not limited the enforceability of their aggressive terms (aggressive at least according to E.U. standards) with the exception of Bragg v. Linden Research Inc. (discussed below, section 3). This could explain why virtual world providers do not feel any urge to take state regulation into account. We will see in the fourth part of this paper why this posture should change. 2.2 Real world law The real world law is the second component in virtual worlds regulation. In the context of this paper, this expression is used as a synonym of state regulation, and includes international law, regional/national/federal law, as well as state law. The expression law should be construed as including all forms of law, whether they are issued by a legislative body (i.e. E.U. directives and regulations, and state issued legislation) or by a court of law. State regulation is secondary because, as was shown previously (section 2.1), the primary source of law lies with the EULAs. Real world law is applied to complement the EULA, if an issue is not covered therein (a rare occurrence since virtual worlds providers tend to comprehensively regulate their relationship with customers) or to redress terms which do not comply with state regulation.

Antreasyan

There is no legislation that directly aims at regulating virtual worlds, which is quite understandable given the recent emergence of this area. In the few cases litigated in the U.S. (e.g. the Bragg and the recent Evans v. Linden Research Inc. case), the Court applied general laws to this new phenomenon. Nonetheless, some regulators are well aware of issues pertaining to virtual worlds. By way of example, the Peoples Republic of China imposes a personal income tax of 20% on profits from virtual money (Ye, 2008). Tax concerns were also raised in the U.S. (see the National Taxpayer Advocate's 2008 Annual Report to Congress). In 2008, a report from the ENISA (European Network and Information Security Agency) addressed a fair amount of legal issues originating from virtual worlds (e.g. risks to intellectual property (IP), privacy risks, and problems with online dispute resolution in virtual worlds). 3 PRIVATE REGULATION VERSUS U.S. REGULATION

This section will provide an overview of conflicts between the private regulation of virtual worlds and state regulation in the U.S., relying particularly on two cases: Bragg v. Linden Research and Evans v. Linden Research. Both of these cases relate to the arbitration and the choice-of-forum provisions found in Second Lifes EULA. In Bragg, the court focused primarily on the binding arbitration term, which, at the time of the dispute, compelled users of Second Life to bring their claims to arbitration in San Francisco. The Federal Arbitration Act provides that an arbitration clause shall be valid, subject to generally applicable contract defenses, such as fraud, duress, or unconscionability (9 U.S.C. 2; Doctor's Assocs. v. Casarotto, 517 U.S. 681, 1996). In casu, the issue was whether that binding arbitration term was unconscionable. The Court found that both procedural and substantive unconscionability requirements were met, as required by California law (see e.g. Davis v. O'Melveny & Myers, 2007). First, it was held procedurally unconscionable due to Second Life Terms of Service (ToS, which is interchangeable with EULA) being considered a contract of adhesion (Bragg v. Linden Research, at 605-607). Second, substantive unconscionability focuses on the one-sidedness of contract terms. If a term is so one-sided as to be unfair to the weaker party, such term shall not be enforceable. The court held the arbitration clause substantively unconscionable for several reasons: in particular due to the lack of mutuality (Linden had self-help remedies, e.g. the right to terminate users account without notice, before arbitrating a dispute, whereas Mr. Braggs sole remedy was arbitration), and due to the unreasonable cost of arbitration (i.e. the cost of arbitration itself and the cost linked to the arbitration venue in San Francisco) (see Bragg v. Linden Research, at 607-611; see also Sites et al., 2010). The case was ultimately settled. The substantive claims whether Linden Research had rightfully closed Braggs account and (in effect) extorted all his virtual properties by applying its EULA were thus not addressed by a court of law. Consequently, it remains unclear whether any jurisdiction in the U.S. would hold such terms unconscionable. The same question was brought before the Court in Evans, but the contentious arbitration clause had changed. Linden Research had it amended and modeled after the eBay dispute resolution mechanism: mandatory jurisdiction/venue in Second Lifes home court except for permissive [non-appearance-based] arbitration for [] disputes [where the total amount sought is less than USD 10000.00] (Goldman, 2011; see also Art. 12.1 Second Life ToS). The Court held that this mere change was sufficient to enforce that term, without considering whether mutuality was lacking (Linden Research still has the drastic selfhelp remedy of user account termination, see Art. 11.3 Second Life ToS). The evolution between these two cases, if Evans constitutes a precedent, shows that control over virtual worlds is not evolving toward more protection for consumers in the U.S. On the contrary, the liberal U.S. regulation allows private entities namely virtual world providers to retain a lot of control over users of their service.

Antreasyan

PRIVATE REGULATION VERSUS E.U. REGULATION

After assessing the conceptual importance of virtual world EULAs, and their conflicts with U.S. regulation, this section will focus on EULAs enforceability under E.U. regulation. The first step will be to justify, on the one hand, the application of E.U. law to U.S. contracts (4.1) which is not a trivial issue given that virtual world EULAs generally elect U.S. law to govern the contract and, on the other hand, the jurisdiction of E.U. courts over contractual virtual worlds disputes involving an E.U. resident (4.2). Then, virtual world EULAs will be scrutinized under two important European substantive laws: consumer protection law (4.3), and intellectual property (IP) law (4.4). 4.1 Why would E.U. regulation be applicable to U.S. EULAs ? There has been several discussions about how general real world laws should apply to virtual worlds (see e.g. Duranske, 2008; Lastowka, 2010). However, another interesting question to be answered in this section prior to discussing the limitations that substantive E.U. regulation (or E.U. Member States legislation) could impose to U.S. contracts concerns which real world law is applicable, i.e. the law of which state or region governs the contract? All of the EULAs studied contain a choice-of-law provision. Essentially for financial and practical reasons, virtual world providers usually elect the law of the state where they are headquartered (e.g. the laws of the State of California in the case of Second Life, or the federal laws of the United States of America and the laws of the State of Delaware in the case of World of Warcraft). Therefore, we can validly ask: why would E.U. regulation apply to these agreements and potentially limit their enforceability? In this context, conflict-of-law rules applying to consumer contracts (4.1.1) and IP contracts (4.1.2) will be analyzed because some provisions of these types of contracts may trigger the application of E.U. mandatory rules, which would render such provisions unenforceable (see section 4.3 and 4.4). 4.1.1 The law applicable to consumer contracts According to Article 6.1 of Council Regulation (EC) No 593/2008 on the law applicable to contractual obligations (Rome I Regulation), consumer contracts
shall be governed by the law of the country where the consumer has his habitual residence, provided that the professional [], by any means, directs such activities to that country or to several countries including that country, and the contract falls within the scope of such activities.

Article 6 of the Rome I Regulation defines the consumer as a natural person, who enters into a contract for a purpose which can be regarded as being outside his trade or profession. This negative definition is widely accepted in E.U. Member States legal system (Tang, 2009). Additionally, some authors have argued that it should be interpreted narrowly (Garcia, 2005; Piedelivre, 2008; Tang, 2009). In Johann Gruber v Bay Wa AG, the European Court of Justice (ECJ) said:
[A] person who concludes a contract for goods [or services] intended for purposes which are in part within and in part outside his trade or profession may not [be considered as a consumer], unless the trade or professional purpose is so limited as to be negligible in the overall context of the supply, the fact that the private element is predominant being irrelevant in that respect ( 54).

Accordingly, it could be argued that users who conduct business in virtual worlds should not be regarded as consumers they need to access the service (and thus to enter into the EULA) to pursue their trade or profession. The rationale of Article 6 of the Rome I Regulation is to protect the weaker party who is not a professional (Ragno, 2009; Tang, 2009), i.e. it is assumed that a person acting as a professional cannot be weaker. However, in case of an agreement between a virtual world provider and its user, the latter is substantially weaker than the former, even the very few users who make a living in virtual worlds. It could thus validly be argued that at least partly professional users of virtual worlds should benefit from the special protection of this provision (supporting the view that the predominance test should be
Antreasyan 4

applied: Giuliano/Lagarde, 1980). Regarding users who conduct virtual business in addition to a real world trade or profession (e.g. Mr. Bragg who traded virtual goods in Second Life and is an attorney in the real world), we could argue that even if they made a profit in the virtual world, it is precisely made outside of their real trade or profession, thus qualifying them as consumers. In any case, the clear wording of Article 6.1 should prevent a company which is not a natural person as well as a natural person who accesses a virtual world with the main purpose of conducting business, from invoking this protective statute. Another condition for Article 6.1 to be applicable is that the provider shall direct its activities toward a Member state. In that respect, Recital 24 of the Rome I Regulation states (citing a joint declaration by the Council and the Commission on Article 15 of Council Regulation (EC) No 44/2001 (Brussels I Regulation)):
[T]he mere fact that an Internet site is accessible is not sufficient for Article [6] to be applicable, although a factor will be that this Internet site solicits the conclusion of distance contracts and that a contract has actually been concluded at a distance, by whatever means. In this respect, the language or currency which a website uses does not constitute a relevant factor.

Although the wording of Recital 24 seems to suggest that the interpretation of directed activities should be narrow, legal scholars recommend interpreting it as broadly as possible (see e.g. De Meyer, 2005; Nielsen, 2007; Ragno, 2009). Nielsen (2007) goes so far as to state that a website, which uses the English language and solicits the conclusion of contracts with consumers, should be deemed to direct its activities to every country in the world. However, the provider can take steps to prevent that outcome, e.g. block websites access for certain countries or exclude consumers from certain countries to enter into the EULA (Nielsen, 2007). Both of these measures are technical. It should thus be asked whether the direction of activities could be excluded through a EULA provision? It should not. For example, Art. 13.1 of Second Lifes Terms of Service states that Linden Lab makes no representation that any aspect of the Service is appropriate or available for use outside of the United States. In spite of that term, the fact is that its service is available (and widely used) outside of the U.S. If including such a term in a EULA would suffice to prevent the application of Article 6 of the Rome I Regulation, it would be contrary to Article 6s purpose (Magnus/Mankowski, 2004). Accordingly, only technical measures should be deemed satisfactory to find that activities of virtual world providers are not directed toward the E.U. Three considerations should be pointed out as regards the law applicable to consumer contracts. First, even if a choice-of-law provision may still be valid if it complies with Article 6.1 (according to Art. 3 and 6.2), EULAs cannot derogate from domestic mandatory rules of the consumers habitual residence (Art. 6.2) (see Tang, 2009). Second, the Council Directive 93/13/EEC of 5 April 1993 on unfair terms in consumer contracts (the Directive) adopts a different conflict-of-law rule based on the close connection between the contract and the Member States (Art. 6.2). However, in the context of virtual world EULAs, this distinction should not have any impact, i.e. if Article 6 of the Rome I Regulation is not applicable, neither will the Directive (on this issue, see Ragno, 2009; and Tang, 2009). Third, it should be noted that Article 6 applies only when a contract is concluded between a professional and a consumer. The latter has been defined above, and the formers definition needs not to be discussed further in context of this paper because it is clear that virtual world providers are professionals. 4.1.2 The law applicable to intellectual property contracts Two situations shall be distinguished in this context: whether users of virtual worlds are considered as consumers on the one hand, and, on the other hand, whether they are not. First, if the user is considered a consumer, it has to be pointed out that since Rome I Regulations entry into force in December 2009 IP contracts are not excluded from the substantive scope of the consumer contracts provision (Art. 6 Rome I Regulation a contrario; see Nishitani, 2009). Thus, contracts entered into with a consumer shall not deprive the consumer of the protection afforded to him by [IP] provisions that cannot be derogated from by agreement by virtue of the law [of the country where the consumer habitually resides] (Art. 6.1 and 6.2 Rome I Regulation).
Antreasyan 5

As a matter of fact, virtual world providers generally tend to regulate IP aspects (e.g. the assignment of rights) in their EULAs, so that they can be regarded as IP contracts. Consequently, the application of E.U. regulation and/or Member States legislation could interfere with the enforceability of such provisions (see section 4.4). Furthermore, the previous development is not constrained to virtual worlds, but could potentially have an impact on all U.S. software or social network contracts; basically impacting all EULAs with IP aspects, when concluded with E.U. consumers. The second situation regards users of virtual worlds that are not considered consumers. In that case, general conflict-of-law principles apply. In accordance with the party autonomy principle (Art. 3 (1) Rome I Regulation), the parties have an extensive freedom to choose the law applicable to the contract (Nishitani, 2009). Nonetheless, it is important to point out, in the context of virtual world EULAs, that Art. 21 Rome I Regulation could restrict the application of the chosen law if such application would be manifestly incompatible with the public policy of the forum. Two elements are noteworthy. First, public policy in the field of IP in Europe can be very restrictive, especially in countries where authors moral rights are distinctively protected (see section 4.4 for some examples). Second, only the public policy provision of the forum state applies, i.e. if the case is litigated before U.S. Courts, E.U. public policy will not be enforced (see Dessemontet, 2004; Nishitani, 2009). It can thus not be stressed enough that providers intending to prevent the application of such mandatory policy must specify a forum state outside the E.U. in their EULA. Most virtual world EULAs include a choice-of-forum term, which usually specifies a forum in the U.S. But some do not. For instance, Booyahs ToS does not include any choice-of-forum provision. 4.2 E.U. courts jurisdiction over contractual related disputes As mentioned above, virtual world EULAs usually designate a forum in the United States in case a dispute arises. However, similar to what was discussed about choice-of-law terms, the E.U. restricts choice-of-forum provisions involving consumers, hence potentially hindering their enforceability (on the definition of consumer, see section 4.1.1). Other grounds for E.U. courts jurisdiction will not be discussed hereafter, as the freedom of contract doctrine allows parties, in principle (Art. 23-24 Brussels I Regulation), to select whatever forum they like. It should be noted however that Article 22 Brussels I Regulation (listing exclusive jurisdictions which shall not be derogated from) sets the forum in the Member State where the IPR is registered in disputes pertaining to the existence and validity of IPR required to be deposited or registered. The Brussels I Regulation provides that choice-of-forum terms in contracts entered into with consumers are invalid unless: (1) the contract was entered into after the dispute has arisen; or (2) if the consumer can still bring proceedings before courts of his habitual residence; or (3) if the selected jurisdiction is also the habitual residence of the business and the consumer at the time of entering into the agreement (Art. 17 Brussels I Regulation; see also Tang, 2009). Consequently, in cases where virtual world EULAs select exclusively a U.S. forum and the consumer resides in an E.U. Member State, the choice-of-forum term will not be valid. As a result of the invalidity of the choice-of-forum term, the default jurisdiction, provided for in Section 4 Brussels I Regulation, would apply. On the one hand, a consumer can bring proceedings against the other party in the country where he is domiciled or where the other party is domiciled (Art. 16.1). On the other hand, the other party (i.e. the virtual world provider) may only bring legal action against a consumer in the Member State where the consumer is domiciled (Art. 16.2). This provision applies only if the provider directs its activities to that Member State (Art. 15). The direction of activities, in Art. 15 Brussels I Regulation and Art. 6 Rome I Regulation, shall be construed uniformly (Recital 24 Rome I Regulation). It can thus be assumed, as seen in section 4.1.1, that virtual world providers from the U.S. direct their activities toward Europe, provided they do not technically prevent European users from accessing their service.

Antreasyan

4.3 Consumer protection law This substantive E.U. regulation has the potential to considerably affect the enforceability of U.S. virtual world EULAs for E.U. consumers. That stems from the (very) protective provisions especially from a U.S. point of view that safeguard consumers interests in Europe (on a brief history of E.U. consumer protection law, see Quinn, 2010). As Winn & Webber (2006) put it, [E.U.] regulators have been expanding their oversight of consumer markets and expanding the role of administrative agencies in enforcement, at precisely the time that U.S. contract law has turned away from public regulatory models. Several Council Directives aim at the protection of consumers, but the Directive on unfair terms in consumer contracts is particularly relevant, and thus the only one that will be discussed in this paper. It has to be noted, for practitioners not familiar with the E.U. legal system, that Council Directives are usually not self-executing. Member States must implement them into their national legislation before they become binding for natural or legal persons (Art. 288 of the Treaty on the Functioning of the European Union). Moreover, this particular Directive is a minimum standard of protection, i.e. Member States may adopt more stringent provisions to ensure a maximum degree of protection for consumers (Art. 8 Directive; see also Stuyck, 2009). After evaluating its personal scope of protection (4.3.1), the objective scope (i.e. unfair terms definition) of the Directive will be presented (4.3.2). Then, specific provisions will be evaluated in relation to the unfair terms provision, with reference to U.S. case law where available (4.3.3). Finally, it will be important to briefly refer to the future of E.U. consumer protection, given that its reconsideration is an important work in progress (4.3.4). 4.3.1 Personal scope of protection of the Directive Following what was introduced in section 4.1.1, the benefit of consumer protection regulations in the E.U. is logically limited to consumers, i.e. natural persons who do not enter into contracts predominantly for professional purposes (see Nebbia, 2010). Under Art. 2 of the Directive, the consumer should be understood in the same way as in Art. 6 Rome I Regulation (see Nebbia, 2010). However, since the Directive is a minimum standard of protection, E.U. Member States could adopt a wider definition of consumers. In France for instance, legal persons can be considered consumers (Piedelivre, 2008). 4.3.2 Subject matter of the Directive: what are unfair terms? Unfair terms are defined at Article 3 of the Directive, which states:
1. A contractual term which has not been individually negotiated shall be regarded as unfair if, contrary to the requirement of good faith, it causes a significant imbalance in the parties' rights and obligations arising under the contract, to the detriment of the consumer. [] 3. The Annex shall contain an indicative and non-exhaustive list of the terms which may be regarded as unfair.

Accordingly, two criteria are needed to consider a term unfair: the term must be contrary to good faith and (cumulatively) cause a significant imbalance between parties to the agreement. Some authors debate the good faith requirement on the issue of knowing if it is a procedural or substantive requirement (see e.g. Nebbia, 2010; Stuyck, 2009). In a case brought before the House of Lords (in the U.K.), Lord Bingham of Cornhill explained that the requirement of good faith is one of fair and open dealing. Openness requires that the terms should be expressed fully, clearly and legibly, containing no concealed pitfalls or traps. [] Fair dealing requires that a supplier should not, whether deliberately or unconsciously, take advantage of [e.g.] the consumer's necessity, indigence, lack of experience, unfamiliarity with the subject matter of the contract, weak bargaining position (Director

Antreasyan

General of Fair Trading v. First National Bank Plc). That definition of good faith clearly supports the idea that the good faith criterion is procedural. This is convincing because if it were to be understood as a substantive requirement it would merge with the significant imbalance requirement (Nebbia, 2010). Moreover, Nebbia (2010) states that the assessment of good faith would include verifying whether the consumer had the chance to influence the terms of the agreement, whether there was a choice in agreeing to the terms or if there were alternatives. Users of virtual worlds have to accept EULAs on a take-it-or-leave-it basis, and thus have no chance of negotiating with the provider. The question of alternatives or the absence of choice can be left unanswered in this context, as most virtual worlds would not pass the negotiation test. It is interesting to note that the majority of E.U. Member States did not implement the good faith criterion (Stuyck, 2009), deriving the unfairness of a term solely from the significant imbalance test. As this kind of implementation is more consumer-friendly, it would not be contrary to the Directive which constitutes a minimum standard of protection toward consumers. The significant imbalance criterion involves a lack of symmetry in parties rights and obligations, or that the sellers or suppliers rights or remedies are excessive and disproportionate (Nebbia, 2010). This criterion cannot be evaluated in abstracto, as opposed to the good faith requirement (that is in the context of virtual world EULAs). Widespread terms in virtual world EULAs will precisely be confronted to this test in the next section (4.3.3). Furthermore, the assessment of unfairness shall be made by referring, at the time of conclusion of the contract, to all the circumstances attending the conclusion of the contract and to all the other terms of the contract (Art. 4.1 Directive). The assessment of the unfairness shall not relate to the contracts subject matter or to the adequacy of the price or remuneration (Art. 4.2 Directive). In addition, according to Article 5 of the Directive, terms shall be written in plain, intelligible language, which is a component of the English good faith requirement (see above). To comply with this requirement Blizzard uses a different EULA for European consumers, which is actually intelligible as opposed to most virtual world EULAs (on the other differences between Blizzards U.S. and E.U. EULAs, see Quinn, 2010). In its implementation, the French legislator provided that terms must be available in the French language (Law No 94-665 of August 4, 1994), which is usually not the case for virtual world EULAs (e.g. Second Life, MyTown, and FarmVille) with Blizzard being the exception (World of Warcraft ToS are available in French at http://eu.blizzard.com/fr-fr/company/legal/wow_tou.html). On a side note, it is interesting to notice that Facebook and Twitter have localized their EULAs. However, they both provide that only the English version shall be enforceable. Whether that would be satisfactory for French courts remains unclear. The Directive provides an annex containing a list of terms which may be regarded as unfair. It has been implemented as a black list in some E.U. Member States, which means that each of the lists terms is deemed unfair. For instance, a term requiring the consumer to take disputes exclusively to arbitration not covered by legal provisions ((q) Annex of the Directive) would be unfair. The ECJ was asked several times to interpret the unfairness of contract terms, but except where issues pertained to choice-of-forum or arbitration terms, it refused to rule, arguing that the interpretation of unfairness lies with national courts (Freiburger Kommunalbauten; see also Nebbia, 2010). Finally, the consequence of the fairness test is that if a term is deemed unfair, it will not be binding on the consumer. Nonetheless, the rest of the contract remains binding if it is capable of continuing in existence without the unfair terms (Art. 6.1 Directive). 4.3.3 The AOL case and endangered EULA terms A significant case brought before the French Court of Appeal of Versailles in 2004 will help to put the last section into perspective. The French consumer protection organization UFC Que Choisir asked the Court to challenge 36 terms of AOLs EULA. The severe verdict held that 31 of those terms were unfair or illegal. For instance: the requirement that users constantly update their data or see their account terminated; the transmission of subscribers data to third parties without their prior consent; the unilateral
Antreasyan 8

right of AOL to modify the agreement; the right of AOL to terminate the account without prior notice for EULA breach; or the non-exclusive assignment to AOL of rights to all content put online by subscribers (see Joslove & Krylov, 2005; Winn & Webber, 2006). Consequently, the Court ordered AOL to publish the judgment on its website and erase the unfair/illegal terms from its EULA. These kinds of terms can be found in many virtual world EULAs, which are thus exposed to the review of E.U. courts. Other terms commonly used in these EULAs will be reviewed hereafter. We have already seen, in sections 4.1 and 4.2, that choice-of-law and choice-of-forum terms were unenforceable against E.U. consumers, where they select a law or forum outside the consumers country of residence. Another classic term, which hinders users access to justice, regards dispute resolution through binding arbitration a question that was adjudicated in the U.S. in Bragg and Evans is invalid in Europe (Quinn, 2010; see also the Directive on consumer rights proposal, 4.3.4). Without assessing the legal nature of virtual property, it can safely be assumed that EULAs, which provide that users do not have any virtual property rights, and in which such property can be extorted from them by the provider (e.g. by terminating their account), could be considered unfair by E.U. courts, especially in consideration of the efforts/investments committed by users. 4.3.4 The Future of E.U. consumer protection regulation In 2008, the E.U. Commission proposed a Directive on Consumer Rights (Proposal). It changes the approach of consumer protection regulation in Europe, shifting it from several minimum standard Directives to one comprehensive Directive on consumer rights which shall be transposed as is in all E.U. Member States (Tichy, 2010). In the majority of E.U. Member States the Proposal would increase consumer protection, but in some countries, like Germany or Finland, the standard of protection would be reduced (Brichacek, 2010). The Proposal introduces a new way of controlling unfair terms, which are classified in three categories: the black list includes terms considered unfair in all circumstances, e.g. mandatory arbitration (Art. 34); the grey list includes terms presumed to be unfair (Art. 35); and Art. 32 provides general principles to assess the unfairness of terms (which are not listed in any category). Furthermore, the Proposal reaffirms, in a more explicit way, means of enforcement in relation to unfair contract terms. In this respect, Article 38.3 states: Member States shall enable the courts or administrative authorities to apply appropriate and effective means to prevent traders from continuing to use terms which have been found unfair. This provision could have disastrous effects for providers that did not draft their EULAs in compliance with E.U. regulation (see the AOL case, section 4.3.3). 4.4 Intellectual property law As seen above, E.U. public policy pertaining to IPR may apply in some circumstances (4.1.2), thus interfering with EULA provisions. Selected examples of such policy found in E.U. Member States will be briefly presented below. Some thoughts about what could be called intellectual property private policy efforts will then be offered. Virtual world EULAs routinely address IPR issues pertaining to user-generated content, but in various ways. For example, Second Life is famous for leaving all IPR for such content to its creator (Art. 7.1 Second Life ToS). Other providers, such as Blizzard, grant themselves ownership of all IPR relating to the virtual world (Ownership section in World of Warcraft ToS). This last kind of term could prove troublesome under certain European countries legislation on copyright. For instance, a general transfer of IPR shall not be permitted in France (Art. L-131.3 of the French Code of Intellectual Property). Germany has a more rigid approach: only the execution of IPR shall be transferable, authors rights are not (29 German Copyright Law). Moreover, the European Copyright Code which is a restatement of law project drafted by European IP scholars states: [t]he author can consent not to exercise his moral rights. Such consent must be limited in scope [i.e. general waivers are not possible, but an author may consent to particular uses], unequivocal and informed.
Antreasyan 9

Following some concerns about how they handle IP infringement on their websites/virtual worlds, some providers have implemented IP policies. This is the case of Facebook, Twitter, and more interestingly in the context of this paper, Second Life. The latter put the Digital Millenium Copyright Act (DMCA) into effect and, additionally, created a trademark infringement notification system (Linden Lab Official: Intellectual Property). Interestingly, this system benefits only owners of U.S. trademarks, thus discriminating against international trademark owners. Facebooks policy is similar and it has raised concerns at the World Intellectual Property Organization (WIPO): [f]or example, the current form for reporting an infringing username requires a trademark registration number. Moreover, the form states a preference for US registration numbers. It is unclear whether the procedure is adequately equipped to deal with territorial conflicts, complaints of brand owners lacking a US registration, or brand owners simply asserting common law rights in their marks. (WIPO, Trademarks and the Internet). 5 CONCLUSION

In the battle between virtual world providers and their users, the stronger opponent clearly is the provider. The U.S. referee does not have a substantial influence over the game, and the E.U. referee, who could potentially have much more influence, did not use its whistle once in a virtual world context. However, providers are not immune from a U.S. policy change and more importantly at least in the foreseeable future from litigations initiated, in the E.U., by European consumers (or consumer organizations). The latter could have immense effects on virtual worlds regulation. Courts applying E.U. law would in fact declare most terms of EULAs unfair and thus would not enforce them, actually rendering the private regulation of virtual worlds ineffective. So what should U.S. providers do, provided they do not want to stay away from the European market? There are several possible course of action: (1) adopt a wait-and-see approach (i.e. keep using an aggressive EULA and probably face E.U. disputes in a relatively near future) and amend their EULA on a case-by-case basis (e.g. Linden Lab); (2) use different EULAs depending on the country/region where users are located (e.g. Blizzard); or (3) use one EULA for all users around the world, which would be drafted according to the most consumer-friendly standards (currently found in the E.U.) (see Quinn, 2010). The first strategy is unpredictable, particularly in view of the stringent consumers protection in Europe. The second and third approaches appear more cautious. To assess which one to choose, virtual world providers should evaluate the cost of each approach. Exactly how that cost would be estimated is beyond the scope of this paper, but a few elements could be mentioned: drafting costs, organizational costs, and litigation costs (which could increase or decrease in the E.U. or in the U.S., depending on the chosen solution). At the very least, virtual world providers should consider the consequences of their current EULA strategy: one should know the rules before playing the game.

Antreasyan

10

BIBLIOGRAPHY Literature
Brichacek, T. (2010). Some Remarks on the Proposal of Directive on Consumer Rights. In Hans Schulte-Nlke and Tichy Lubos (Eds.) Perspectives for European Consumer Law - Towards a Directive on Consumer Rights and Beyond. Munich: Sellier, 97-98. De Meyer, J. (2005). International jurisdiction and conflict of law rules for consumer claims: a survey of European legislation. Revue europenne de droit de la consommation, 631-677. Dessemontet, F. (2004). Copyright contracts and choice of law. In Nordemann Wilhelm (Ed.) Urheberrecht im Informationszeitalter: Festschrift fr Wilhelm Nordemann zum 70. Geburtstag am 8. Januar 2004. Munich: Verlag C.H. Beck, 415-427. Available at: http://www.unil.ch/webdav/site/cedidac/shared/Articles/Mlanges %20Nordemann%20.pdf. Duranske, B. (2008). Virtual Law: Navigating the Legal Landscape of Virtual Worlds. Chicago, IL, USA: American Bar Association. Fairfield, J. A. T. (2008). Anti-Social Contracts: The Contractual Governance of Virtual Worlds. McGill Law Journal, 427-476. Available at: http://lawjournal.mcgill.ca/documents/Fairfield.pdf. Garcia, K. (2005). La dfinition du consommateur et la qualification des actions contractuelles et non contractuelles. Revue europenne de droit de la consommation, 617-630. Giuliano, M. and Lagarde, P. (1980). Report on the Convention on the law applicable to contractual obligations. Official Journal of the European Communities, No C 282/1, 1-50. Available at: http://aei.pitt.edu/1891/1/ Obligations_report__Guiliano_OJ_C_282.pdf. Jankowich, A. (2006). Eulaw: The Complex Web Of Corporate Rule-Making In Virtual Worlds. Tulane Journal of Technology and Intellectual Property, 8, 1-58. Joslove, B. and Krylov, A. (2005). Standard American Business to Consumer Terms and Conditions in the EU. Mich. Intl Law, 18, 1-4. Available at http://www.michbar.org/international/pdfs/Spring05.pdf. Lastowka, G. F. (2010). Virtual Justice: The New Laws of Online Worlds. New Haven, CT: Yale University Press. Lessig, L. (2008). Code: And Other Laws of Cyberspace, Version 2.0. New York, NY: Basic Books. Magnus, U. and Mankowski, P. (2004). The Green Paper on a Future Rome I Regulation-on the Road to a Renewed European Private International Law of Contracts. 103 Zeitschrift fr Vergleichende Rechtswissenschaft 131. Nebbia, P. (2010). Unfair Contract Terms. In Christian Twigg-flesner (Ed.) The Cambridge Companion to European Union Private Law. Cambridge: Cambridge University Press, 216-228. Nielsen, P. A. (2007). Art. 15-17. In Ulrich Magnus and Peter Mankowski (Eds.) Brussels I Regulation. Munich: Sellier, 306-325. Nishitani, Y. (2009). Contracts Concerning Intellectual Property Rights. In Franco Ferrari and Stefan Leible (Eds.) Rome I Regulation - The Law Applicable to Contractual Obligations in Europe. Munich: Sellier, 51-84. Piedelivre, S. (2008). Droit de la consommation. Paris: Economica. Quinn, P. J. (2010). A Click Too Far: The Difficulty In Using Adhesive American Law License Agreements To Govern Global Virtual Worlds. Wisconsin International Law Journal, 757-789. Ragno, F. (2009). The Law Applicable to Consumer Contracts under the Rome I Regulation. In Franco Ferrari and Stefan Leible (Eds.) Rome I Regulation - The Law Applicable to Contractual Obligations in Europe. Munich: Sellier, 129-170. Risch, M. (2009). Virtual Rule of Law. West Virginia Law Review, 1, 1-52. Available at: http://papers.ssrn.com/ sol3/papers.cfm?abstract_id=1463583. Sites, B. D., Peele, C. A., and Fairfield, J. A. T. (2010). End-User License Agreements: The Private Law in Video Games and Virtual Worlds. In Ross A. Dannenberg, Steve Mortinger, Roxanne Christ, and Chrissie Scelsi (Eds.) Computer Games and Virtual Worlds: A New Frontier in Intellectual Property Law. Chicago, IL: American Bar Association, 5-46. Stuyck, J. (2009). Unfair Terms. In Geraint Howells & Reiner Schultze (Eds.) Modernising and Harmonising Consumer Contract Law. Munich: Sellier, 115-144.

Antreasyan

11

Tang, Z. S. (2009). Electronic Consumer Contracts in the Conflict of Laws (Studies in Private International Law). Oxford: Hart Publishing. Winn, J. K., & Webber, M. (2006). The Impact Of EU Unfair Contract Terms Law On U.S. Business-To-Consumer Internet Merchants. Business Lawyer, 62, 209-228. Available at: http://www.law.washington.edu/Directory /docs/Winn/winn_webber_unfair.pdf.

Cases
AOL France v. UFC Que Choisir, R.G. No 04/05564, Cour d'appel de (Court of Appeals of) Versailles, Sep. 15, 2005. Available at: http://www.juriscom.net/documents/caversailles20050915.pdf. Bragg v. Linden Research Inc., 487 F. Supp. 2d 593 (E.D.Penn. 2007). Davis v. O'Melveny & Myers, 485 F.3d 1066 (9th Cir. 2007). Director General of Fair Trading v. First National Bank Plc [2001] U.K.H.L. 52. Available at: http://www. publications.parliament.uk/pa/ld200102/ldjudgmt/jd011025/fair-1.htm Doctor's Assocs. v. Casarotto, 517 U.S. 681 (U.S. 1996). Evans v. Linden Research Inc. case (E.D. Pa. Feb. 3, 2011). Freiburger Kommunalbauten v. Hofstetter (Case C-237/02, Apr. 1, 2004). Johann Gruber v. Bay Wa AG (Case C-464/01, Jan. 20, 2005).

Documents
FarmVille (Zynga) Terms of Service (November 30, 2010). Available at: http://www.zynga.com/about/terms-ofservice.php. MyTown (Booyah) Terms of Service (October 6, 2010). Available at: http://www.booyah.com/terms/. Linden Lab Official: Intellectual Property. Available at: http://wiki.secondlife.com/wiki/Linden_Lab _Official:Intellectual_Property. National Taxpayer Advocate's 2008 Annual Report to Congress. Department of the Treasury, Internal Revenue Service. Available at: http://www.irs.gov/advocate/article/0,,id=202276,00.html. Second Life (Linden Lab) Terms of Service (December 15, 2010). Available at: http://secondlife.com/ corporate/tos.php. Virtual Worlds, Real Money Security and Privacy in Massively-Multiplayer Online Games and Social and Corporate Virtual Worlds (2008). European Network and Information Security Agency (ENISA). Available at: http://www.enisa.europa.eu/media/press-releases/2008-prs/virtual-worlds-real-money. WIPO, Trademarks and the Internet (2010), Standing Committee on the Law of Trademarks, Industrial Designs and Geographical Indications, SCT/24/4. Available at: http://www.wipo.int/edocs/mdocs/sct/en/sct_24/ sct_24_4.pdf. World of Warcraft (Blizzard) U.S. Terms of Use Agreement (December 9, 2010). Available at: http://us.blizzard. com/en-us/company/legal/wow_tou.html. World of Warcraft (Blizzard) E.U. Terms of Use Agreement (December 9, 2010). Available at: http://eu.blizzard. com/en-gb/company/legal/wow_tou.html or in French (December 14, 2010) http://eu.blizzard.com/frfr/company/legal/wow_tou.html.

Newspaper and website articles


Goldman, E. (February 9, 2011). Second Life Forum Selection Clause Upheld Evans v. Linden. Available at: http://blog.ericgoldman.org/archives/2011/02/second_life_for.htm. Ye, J. (October 31, 2008). Real Taxes for Real Money Made by Online Game Players. Wall Street Journal. Available at: http://blogs.wsj.com/chinarealtime/2008/10/31/real-taxes-for-real-money-made-by-online-gameplayers/.

Antreasyan

12

The Game Behind the Video Game: Business, Regulation, and Society in the Gaming Industry New Brunswick, New Jersey, USA, April 8-9, 2011

Copyrights Middle Ground: The Interaction of Players and Video Games


Bruce E. Boyden Marquette University Law School P.O. Box 1881, Milwaukee, WI 53201-1881, USA

ABSTRACT What is copyrightable about a video game? Copyright law treats the appearance and sound of a video game in play as an audiovisual work, just like films and television programs. This categorization fails to recognize the unique characteristics of video games and has some potentially deleterious consequences for users. Among other things, it treats a players recording of their own game play as a verbatim reproduction of the worksomething that would have difficulty qualifying as a fair use. By eliding player contributions to the game in play, copyright avoids doctrinal complexity at the expense of overprotection. Instead, copyright law should apply a different rule to video games, one that originated in the context of ordinary card and board games: namely, that games are not copyrightable. Copyright has long excluded games from its protection, even though each of the component elements of a gamerule sheet, pieces, board, cards, etc.can be protected. Although the original purpose of the rule is obscure, it has the salubrious effect of insulating the activity of play from the reach of copyright. That limit on copyright is important because game play represents a middle ground between the authors expression and a users use, one in which the users perceptual experience of the work melds with the work as it unfolds. The middle ground is where a user actually experiences the workfor example, in reading a book, or watching a film. Like thought, play is free from liability. The result is that while a video games surface elements are copyrightable, the second-order design that gives rise to play, such as intriguing challenges or novel traps, are not. This limit on game copyrights also has implications for copyright law generally. KEYWORDS: copyright, games, video games, players, interactivity, system 1 INTRODUCTION: MODERN VIDEO GAMES AND THE PROBLEM OF PLAYER INTERACTION

In the beginning there were only books. Copyright law as originally formulated protected books from being reprinted without the authors permission. But by the end of the nineteenth century, its logic had been extended to creative endeavors in several other media: maps, charts, musical compositions, plays, translations, paintings, drawings, statues, and photographs. By the adoption of the Copyright Act of 1976, the process was complete; under Section 102(a) of that Act, all works of authorship fixed in a tangible medium of expression are protectable under copyright law, regardless of what form the authors creativity takes. That extension of form, however, has revealed a structural weakness at the heart of copyright law. When copyright was restricted to a limited number of well-understood cultural forms, there was little

Boyden

need to identify precisely in the law what it was in those works that was protected. But as the number of forms increased, it became more difficult for courts and legal scholars to identify the acts of authorship in each, the creative expression. Photographs were an early example. Courts continue to struggle with photographs, as they more saliently than other media present a troubling mixture of fact and fiction, of reality and expression. (Farley, 2004). Courts ultimately resolved the issue by scrutinizing the act of taking the photograph in order to locate the discrete creative elements in its creation. (Burrow-Giles Lithographic v. Sarony, 1884; Mannion v. Coors Brewing, 2006) Video games occupy a similar place in the copyright universe. They contain aspects that are uncontroversially copyrightable, such as software code and the graphical elements that appear on the screen during play. But video games in play contain something else: the contributions of the player. Since those player contributions cannot be part of the video game copyright, video games, like photographs, pose challenges for traditional copyright doctrine. But whereas the problems for photographs lie in the mix of elements at the moment of creation, the problems with video games lie at the other end of the process: in the mix of activities at the moment of perception. Photographs pose challenges about authorship; video games pose challenges about readership. Ironically, the case law involving pre-electronic games at some level recognized this difficulty and adopted a rule to address it. It has been blackletter law for decades that games are not copyrightable. (Nimmer & Nimmer, 2010, 2.18[H][3][a]) The purpose of this rule has been somewhat obscured by time, but I have recently argued that it serves the function of separating the copyrightable elements of the game, the game makers creative expression, from the activity of players in playing the game. (Boyden, 2011) That is, the rule insulates the activity of game play from the reach of copyright, because the activity of game play represents the audiences encounter with the work, in the same way that reading a book or listening to music is beyond the reach of copyright. That rule has yet to be applied to video games, however. Instead, early in the history of video games, it was decided that such games are essentially audiovisual works just like cartoons, completely discounting the role of the player in shaping the game display. Those decisions involved early arcade games such as Defender (Williams Elecs. v. Artic Int'l, 1982), Pac-Man (Atari v. N. Am. Philips, 1982; Atari v. Williams, 1981), Galaxian (Midway Mfg. v. Artic Int'l, 1983), Asteroids (Atari v. Amusement World, 1981), and Centipede (Atari v. Armenia, 1981). In todays games, by contrast, the player often has unfettered control over the camera angle, and substantial control over the pace and sequence of events. The player, in other words, could be considered the director of the movie that appears on the screen. Nevertheless, under current law strictly applied, game designers own the audiovisual display of the game each time it is played, which would mean that a player recording his or her own performance is engaged in a verbatim reproduction of the audiovisual work, a strike against any fair use defense. Those early decisions have not been revisited by either courts or scholars. Thomas Hemnes (1982, p. 174) and William Patry (1983, p. 52), for example, accepted the classification of games as audiovisual works and rejected the challenges posed by game play, with Hemnes concluding that such questions were of academic interest only (p. 181). Nimmer has suggested only that the blanket rule of exclusion for games must be rethought as must so much else in the copyright arena insofar as it applies to works of technology heralded by the computer revolution. (Nimmer & Nimmer, 2010, 2.18[H][3][a]) Although there is now a voluminous legal literature on video games, that scholarship has largely focused on other issues, such as virtual property rights (Fairfield, 2005; Lastowka & Hunter, 2004), the relationship between players and game developers (Balkin, 2004; Burk, 2005; Fairfield, 2009), or enduser modifications (Ochoa, 2004). An exception is W. Joss Nichols (Nichols, 2007), who has questioned the continued vitality of the old arcade-game cases, but in the context of an argument that players own the copyright in video game play.

Boyden

As both video games and user-generated content become increasingly important economically,1 courts will no longer be able to avoid determining the scope of video game copyrights. Legal controversies are certain to arise in which players are not only recording or transmitting themselves playing, but that activity is being commercialized in some way. (See, e.g., Marvel Enterprises, Inc. v. NCSoft Corp., 2005) In this paper I argue that the rule against the copyrightability of games ought to be applied to video games as well. That is, video games are not audiovisual works; rather, they are systems, just like other games. Second, I examine what is left of video game copyrights once the uncopyrightable system is excluded. Finally, I conclude by considering the implications for copyright law. 2 COPYRIGHT PROTECTION FOR GAMES

First, a brief synopsis of basic copyright law. Copyright protects creative expression in works such as novels, films, music, newspapers, computer software, paintings, and more. There are three prerequisites to obtaining a valid copyright under the modern Copyright Act: a work must be original to the author, it must be minimally creative, and it must be fixed in any tangible medium of expression, that is, recorded or written down in some fashion. (Copyright Act, 1976, 102(a)) (Registration with the Copyright Office and the inclusion of a copyright notice on the work are no longer required.) Copyright protects an authors expression against not only verbatim duplication by others without permission, but also against later works that are substantially similar to the original. There are several important limits on the extent of copyright protection, however. One is the doctrine that only expression is protected, not general plot concepts or ideas. Second, there are a number of exclusions meant to draw a line between copyright and patent, which has much stricter requirements. Thus, processes, procedures, methods of operation, and systems are not copyrightable, even though descriptions or illustrations of those things might be. (Baker v. Selden, 1879; Samuelson, 2007) Third, there is the doctrine of fair use, under which a use can be made that would otherwise be infringing if it does not sufficiently intrude on the copyright owners work. There a number of other limiting doctrines in copyright law as well, one of which will concern us here: the rule that games are not copyrightable. It is well-established law that games are not copyrightable. (Nimmer & Nimmer, 2010, 2.18[H][3][a]) The rule has long been a puzzling one, however, because courts rarely stated the reasons for it, and because it is fairly limited in scope: while games themselves are not copyrightable, all of their individual components are, such as boards, pieces, cards, and even a particular description of the rules. The explanation that is usually given is that the essence of a game is its rules, and rules are uncopyrightable instructions. The explanation does not withstand scrutiny, however. Game rules are not instructions in the sense that is excluded from copyright; unlike instructions to build a house or make a cake, game rules do not specify the steps necessary to achieve an end result. Rather, they specify a set of constraints and affordances, a starting position, and end conditions. As I have argued previously, that makes game rules similar to another poorly understood exception from copyrightability: systems. (Boyden, 2011, pp. 457-58) The idea that games are systems has also been embraced by game scholars such as Chris Crawford (1984, p. 4), Katie Salen and Eric Zimmerman (2004, p. 50), and Jesper Juul (2005b, p. 36). A system as used in this context is what in computer science is called a state machine: it is a means of transforming a large but constrained set of inputs into a correlated and defined set of outputs. (Juul, 2005b, p. 60) When you make a move in a game, the rules determine how the game state changes in response, and some or all of that information is output back to the players. Systems of this sort are excluded from copyrightability because they do not
At least one report concludes that in 2007 video games reaped $41.9 billion in sales worldwide (Caron, 2008), compared to approximately $40 billion for music, $27 billion for films, and $23 billion for DVDs in the U.S. (Video Game Industry, n.d.)
Boyden 3
1

primarily express ideas to users, but rather serve as empty vessels into which the users pour meaning. In the case of games, that meaning is in the form of the choices made by players. It is important to reiterate that the rule against the copyrightability of games does not leave games unprotected. All of the designed elements of a game can be copyrighted. But the game itself, stripped of all of its ornamentation, cannot. This has two consequences of practical significance. First, a game that plays like an earlier game, but does not duplicate any of the earlier games ostensible features, does not infringe. Second, because the game itself is not copyrightable, playing a game is not the performance of a copyrightable work. (Allen v. Academic Games League, 1996, p. 616) Playing a game in public, therefore, is not an infringing public performance. 3 VIDEO GAMES ARE UNCOPYRIGHTABLE SYSTEMS

Should the rule against the copyrightability of games apply to video games? Courts and commentators have supplied two reasons opposing such an outcome. First, courts have held since the early days of arcade games that video games should be treated as ordinary audiovisual works, not systems or even games. Second, the scholarship on video games suggests an alternative argument: that video games are importantly different from nonelectronic games, in that they contain far more expressive content. 3.1 Video Games as Audiovisual Works Courts first confronted video games in the early 1980s, shortly after video arcade games became a commercial and social phenomenon. The financial success of early arcade games such as Pac-Man, Defender, Asteroids, Space Invaders, and Galaxian generated a wave of knock-offs or close copies by manufacturers attempting to share some of the game-makers good fortune. The infringers were emboldened by not only the gold-rush atmosphere of the market, but also by a state of considerable uncertainty at the Copyright Office and elsewhere over whether video games and computer programs were protectable at all. Lawsuits followed. In those early cases, several defendants tried to challenge the validity of the plaintiffs copyrights in their video games, in particular in the audiovisual display of the game in play. The defendants tended to make two related arguments. First, they argued that the audiovisual display of the game in play failed the originality requirement, because it was not wholly original to the author of the game. Instead, what appeared on the screen was to some degree under the control of the player, not the author.2 Second, they argued that for similar reasons the game in play was not fixed, that is, recordedeach game was different, and none of those games could possibly have been in the version that the plaintiffs deposited with the Copyright Office when they registered their copyrights. The leading case of Stern Electronics, Inc. v. Kaufman (1982) is typical. Stern involved the early 80s arcade game, Scramble, a side-scroller in which the player maneuvered a space ship over mountains while trying to destroy targets. (Fig. 1) The defendant in Stern was unsympathetic; it had produced a blatant Figure 1: Scramble in play mode
This argument received some support in the Copyright Office. In reviewing whether the arcade game Breakout should be registered, the Office initially concluded that the images . . . created by playing the video game . . . are also not registrable since they are created randomly by the player and not by the author of the video game. (Atari Games Corp. v. Oman, 1989, p. 880)
Boyden 4
2

knock-off of Scramble by duplicating the entire game, right down to the cabinet (which bore the same name) and the ROM boards. And the appeals court had little trouble rejecting the defendants arguments. The Stern court held that the amount of variation between different sessions of the game did not undermine the copyright, as many aspects of the sights and the sequence of their appearance remain constant during each play of the game. (p. 856) For example, the court noted, the appearance of the players spaceship and enemy craft, the sequence in which the various missile bases, fuel depots, and terrain appeared, and the sounds that occurred whenever items were destroyed all remained constant. That was sufficient, according to the court, for Scramble to qualify for copyright protection as an audiovisual work. (p. 856) In fact, the Stern court concluded that video games are essentially cartoons in which some of the action is controlled by the player. (p. 853) The court in Midway Manufacturing v. Artic International (1983), reviewing a case involving Galaxian and Pac-Man, reached a similar conclusion: Playing a video game is more like changing channels on a television than it is like writing a novel or painting a picture. The player of a video game . . . cannot create any sequence he wants out of the images stored on the game's circuit boards. The most he can do is choose one of the limited number of sequences the game allows him to choose. He is unlike a writer or a painter because the video game in effect writes the sentences and paints the painting for him; he merely chooses one of the sentences stored in its memory, one of the paintings stored in its collection. (p. 1012) Whatever the validity of these opinions for the early arcade games before the courts at the time, they have been undermined by subsequent developments. Modern games afford the player far more control over the sights and sounds that appear during game play. In many modern games, the player has complete control over the camera, choosing where to look and for how long. The player also has considerable control over the sequence in which events occur. There are obviously limits to this freedom. There are typically boundaries to where a player can travel, and there are usually a limited number of non-player characters in such games. Locations and items tend to look and sound identical to each other. Any in-game dialog or plot is also fixed, in both the sense of being fixed in the code and in the sense of being fixed in a certain time and place in the game. The trend, however, is toward games that are able to generate game environments and action sequences on the fly, in response to a players actions. At some point, the work becomes less of a fixed set of options and more the sort of toolbox envisioned by the Midway court, allowing users to create or act in ways that the author could not have foreseen in detail. Obviously elements of the images and sounds that appear during gameplay are fixed and original. But the overall game as it appears on the screen is not. The influence of the player over the display also undermines the courts classification of videogame displays as audiovisual works. It is no longer necessary to fit copyrightable works into one of the

Figure 2: A crucial moment in System Shock 2, from two perspectives


Boyden 5

eight specifically named categories listed in the Copyright Act. But that classification affects how courts determine what the protectible expression is in that type of work. Courts traditionally look for creativity in audiovisual works by analyzing a discrete set of elements: plot, themes, dialogue, mood, setting, pace, characters, and sequence of events. (Shaw v. Lindheim, 1990, pp. 1356-57). It makes sense to look at these elements in films and television programs because in such works, the precise nature of the display of the work is the result of a series of careful choices by the authors. Directors such as Stanley Kubrick may make dozens of attempts to ensure a single scene occurs exactly as he envisions it, and once the scene has been captured correctly, it will not vary at all from one showing to the next. (See Aarseth, 1997, p. 3) Video games obviously fall short of this standard, for reasons that have often been noted: a players ability to affect the events on the screen is inconsistent with the authorial control behind most audiovisual works. (Fig. 2) Game designers may hope a given scene plays out in a certain way, but in general they determine the average player response and hope for the best. As Salen and Zimmerman (2004) note, [a]s a game designer, you can never directly design play. You can only design the rules that give rise to it. Game designers create experience, but only indirectly. (p. 168) This quickly leads to a thicket of literature, somewhat baffling to an outsider, known as the ludology-narratology debate. It is not the substance of the debate that is confusing, but rather the vehemence with which it is carried out. (Mateas & Stern, 2006) Clearly some games have important narrative aspects to them, others do not, and still others lack any meaningful narrative at all. (Jenkins, 2006, pp. 672-73; Juul, 2005a). For purposes of this paper, none of that matters. Somewhat dangerously, this paper has the potential to bridge the divide by denying both that video games are sufficiently storylike to be copyrightable audiovisual works and that designing for game play is copyrightable expression. This problem of control is reflected in the definition of audiovisual work itself in the Copyright Act: an audiovisual work consists of a series of related images which are intrinsically intended to be shown by the use of machines . . . , together with accompanying sounds, if any. (Copyright Act, 1976, 101) Modern video game displays are not typically, however, a series of images that are permanently embodied (fixed) in the work; rather, they are generated by the program from abstract elements and data in response to the players actions. That is, unlike a film reel or even digital media such as a DVD, there is in a modern video game no fixed set of images to display to the player; there are only the elements of that display, and various algorithms to calculate when and how to arrange them into a display. 3.2 Video Games as Systems It might be objected that even if video games are not audiovisual works, they, unlike nonelectronic games, share a number of features in common with audiovisual works that make application of the traditional rule that games are not copyrightable inappropriate for video games. Video games, for example, often contain plot, themes, dialogue, mood, setting, pace, characters, and a sequence of events. But the rule against the copyrightability of games is not a rule based on the lack of complexity in the expression contained in games. It is rather a rule based on the fact that games are systems, that their purpose is more to provide a forum for experience or information supplied by the user than to convey a particular experience or information from the author to the user. A players play of the game is the way the game is experienced, in the same way that a readers reading of a book is the way that the book is experienced. That act of reading is outside of the scope of the authors copyright in the creative material in the book. Reading is the middle ground between author and reader where the readers thoughts meet the authors expression in a conceptual space that is neither wholly thought nor wholly text.3 Similarly, the act of play is a games middle ground. The difference is that some aspect of that middle ground is visible to third partiesthe players influence over the display of the game. The images and sounds of a game in progress are merely part of the external
The concept of the middle ground between two cultures was developed by historian Richard White (White, 1991, p. ix).
Boyden 6
3

component of a feedback system that is a necessary component of the experience of the game. (Fig. 3) The players thoughts and reactions in response to the game shape his or her operation of the games controls, which in turn affects the computer running the game, which alters the display and sounds, which in turn is perceived by the player, producing new responses. (Salen & Zimmerman, 2004, pp. 315-16)4 While part of this loop is external to the conscious mind, it is inseparable from the players private apprehension of the game, and thus beyond the reach of the game designers copyright.

Figure 3: The video game system


4 IMPLICATIONS FOR VIDEO GAMES

What are the implications of all this for video game copyrights? Does the rule against the copyrightability of games leave video games unprotected? Not at all. Just like nonelectronic games, many of the component elements of a video game audiovisual display can be protected: textures, objects, characters, dialog, plot, scripted sequences, visual effects, sound, lighting, music, and more. All of these elements exist at what Ernest Adams (2009) has called the presentation layer (p. 37); they are directly conveyed to the player. What is uncopyrightable are the design elements that give rise to game play, or the solutions to what Katie Salen and Eric Zimmerman (2004) call a second-order design problem (p. 168). That includes the rules of the video gamethe laws of physics it simulates, the limits on what the player can do, the affordances it givesand also aspects of level design that are intended to shape play. For example, under the theory espoused here, a particularly clever trap is an uncopyrightable gameplay idea, a situation put into the game in order to encourage emergent play, along the same lines as the idea of color groups in Monopoly or triple-word scores in Scrabble. No game designer should have a monopoly over real-estate games with groups of properties or over the prison shoot-out in Half-Life 2. Some other results follow from the argument here. First, because the video game in play is beyond the scope of the copyright, playing a video game is not a performance of a work, which means that playing a video game in public does not infringe on the copyright owners public performance right. Hosting a game server might well be a public performance of the underlying game program, because game play does not directly interact with and alter the copyrighted software, but the display of any particular game session is not a performance. A player recording themselves playing may well copy some
The diagram in Fig. 3 resembles Espen Aarseths (Aarseth, 1997, p. 21) model for all ergodic literature, of operator, text, and medium.
Boyden 7
4

of the protected elements within the presentation layer of the audiovisual display, however. In that case, treating the game display as a collection of discrete elements strengthens the hand of the player, who will likely have a compelling fair use defense to any claim of infringement based in part of the limited nature of copyrightable information captured in a recording and its dissimilarity from the original work, a noninteractive recording as opposed to an interactive game. Second, treating video games as copyrightable only in the elements contained in the presentation layer also helps to analyze those cases where a third party has created some sort of add-on program or level for the game. (Lewis Galoob Toys v. Nintendo, 1992; Micro Star v. Formgen Inc., 1998) Viewing a video game as a sort of audiovisual work encourages courts to find that any alteration of the audiovisual display is an adaptation of the work, requiring a license. Thus the defendants in one case were held liable for selling a collection of add-on levels for Duke Nukem 3D. (Micro Star v. Formgen, 1998). But if what is copyrighted in a video game display is only the individual elements, defendants should only be held liable for infringing adaptations if they copied those elements. The defendants in Micro Star were not, evidently, circulating any of the games original code, and should not have been held liable. 5 CONCLUSION

The rule against the copyrightability of games touches on a deeply significant concept in copyright law, namely that copyright ensures a return on the work of creating and distributing creative expression. Copyright does not reach the act of reading, however. That is, copyrights exclusivity stops at the moment the user encounters the work in order to avoid intruding upon that process. Scholars such as David Nimmer (1996, pp. 42-45) and Jessica Litman (2001, p. 22) have noted the difficulties that would arise if this were not the case: imagining a different ending to a movie might otherwise be infringing. But this limit on copyright has significant consequences beyond the realm of video games. It helps to explain the traditional categories of protection under the fair use defense, for example: criticism, comment, news reporting, teaching . . . , scholarship, or research. (Copyright Act, 1976, 107) All but one of those, news reporting, involve someones attempt to understand and grapple with the meaning of a work, much like play grapples with the meaning of a video game. As copyright law struggles to adapt to the information age, video games will likely pose important test cases in establishing its new boundaries. REFERENCES Aarseth, E. J. (1997). Cybertext: Perspectives on Ergodic Literature. Baltimore: Johns Hopkins University Press. Adams, E. (2009). Fundamentals of Game Design (2d ed.): New Riders Press. Allen v. Academic Games League of America, Inc., 89 F.3d 614 (9th Cir. 1996). Atari Games Corp. v. Oman, 888 F.2d 878 (D.C. Cir. 1989). Atari, Inc. v. Amusement World, Inc., 547 F. Supp. 222 (D. Md. 1981). Atari, Inc. v. Armenia, Ltd., Copy. L. Rep. (CCH) 25,328 (N.D. Ill. Nov. 3, 1981). Atari, Inc. v. N. Am. Philips Consumer Elecs. Corp., 672 F.2d 607 (7th Cir. 1982). Atari, Inc. v. Williams, 217 U.S.P.Q. 746 (E.D. Cal. 1981). Baker v. Selden, 101 U.S. 99 (1879). Balkin, J. M. (2004). Virtual Liberty: Freedom to Design and Freedom to Play in Virtual Worlds. Virginia Law Review, 90(8), 2043-2098. Boyden, B. E. (2011). Games and Other Uncopyrightable Systems. George Mason Law Review, 18(2), 439. Burk, D. L. (2005). Electronic Gaming and the Ethics of Information Ownership. International Review of Information Ethics, 4, 39. Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53 1884). Caron, F. (2008, June 18). Gaming Expected to Be a $68 Billion Business by 2012. Ars Technica. Copyright Act of 1976, 17 U.S.C. 102 (1976). Crawford, C. (1984). The Art of Computer Game Design. Berkeley, CA: Osborne/McGraw-Hill.
Boyden 8

Fairfield, J. A. (2005). Virtual Property. Boston University Law Review, 85, 1047. Fairfield, J. A. (2009). The God Paradox. Boston University Law Review, 89, 1017. Farley, C. H. (2004). The Lingering Effects of Copyright's Response to the Invention of Photography. University of Pittsburgh Law Review, 65, 385. Hemnes, T. M. (1982). The Adaptation of Copyright Law to Video Games. University of Pennsylvania Law Review, 131, 171. Jenkins, H. (2006). Game Design as Narrative Architecture. In K. Salen & E. Zimmerman (Eds.), The Game Design Reader: A Rules of Play Anthology (pp. 670-89). Cambridge: MIT Press. Juul, J. (2005a). Games Telling Stories?. In J. Raessens & J. Goldstein (Eds.), Handbook of Computer Game Studies (pp. 219-26). Cambridge: MIT Press. Juul, J. (2005b). Half-Real: Video Games Between Real Rules and Fictional Worlds. Cambridge: MIT Press. Lastowka, F. G., & Hunter, D. (2004). The Laws of the Virtual Worlds. California Law Review, 92, 1. Lewis Galoob Toys, Inc. v. Nintendo of America, Inc, 964 F.2d 965 (9th Cir. 1992). Litman, J. (2001). Digital Copyright. Amherst, NY: Prometheus Books. Mannion v. Coors Brewing Co., 377 F. Supp. 2d 444 (S.D.N.Y. 2006). Marvel Enterprises, Inc. v. NCSoft Corp., 74 U.S.P.Q.2d 1303 (C.D. Cal. 2005). Mateas, M., & Stern, A. (2006). Interaction and Narrative. In K. Salen & E. Zimmerman (Eds.), The Game Design Reader: A Rules of Play Anthology (pp. 643-69). Cambridge: MIT Press. Micro Star v. Formgen Inc., 154 F.3d 1107 (9th Cir. 1998). Midway Mfg. Co. v. Artic Int'l, Inc., 704 F.2d 1009 (7th Cir. 1983). Nichols, W. J. (2007). Painting Through Pixels: The Case for a Copyright in Videogame Play. Columbia Journal of Law & the Arts, 30, 101. Nimmer, D. (1996). Brains and Other Paraphernalia of the Digital Age. Harvard Journal of Law & Technology, 10, 1. Nimmer, M. B., & Nimmer, D. (2010). Nimmer on Copyright. Ochoa, T. T. (2004). Copyright, Derivative Works, and Fixation: Is Galoob a Mirage, or Does the Form(gen) of the Alleged Derivative Work Matter? Santa Clara Computer & High Tech. Law Journal, 20, 991. Patry, W. (1983). Electronic Audiovisual Games: Navigating the Maze of Copyright. Journal of the Copyright Society of the U.S.A., 31, 1-46. Salen, K., & Zimmerman, E. (2004). Rules of Play: Game Design Fundamentals. Cambridge: MIT Press. Samuelson, P. (2007). Why Copyright Law Excludes Systems and Processes from the Scope of Its Protection. Texas Law Review, 85, 1921. Shaw v. Lindheim, 919 F.2d 1353 (9th Cir. 1990). Stern Elecs., Inc. v. Kaufman, 669 F.2d 852 (2d Cir. 1982). Video Game Industry. (n.d.). Retrieved from the Video Game Industry wiki: http://vgsales.wikia.com/wiki/Video_game_industry. White, R. (1991). The Middle Ground: Indians, Empires, and Republics in the Great Lakes Region, 16501815. New York: Cambridge University Press. Williams Elecs., Inc. v. Artic Int'l, Inc., 685 F.2d 870 (3d Cir. 1982).

Boyden