European Journal of Sport Science

ISSN: 1746-1391 (Print) 1536-7290 (Online) Journal homepage: https://www.tandfonline.com/loi/tejs20

Practice and play in the development of German

top-level professional football players

Manuel Hornig, Friedhelm Aust & Arne Güllich

To cite this article: Manuel Hornig, Friedhelm Aust & Arne Güllich (2016) Practice and play in the
development of German top-level professional football players, European Journal of Sport Science,
16:1, 96-105, DOI: 10.1080/17461391.2014.982204

To link to this article: https://doi.org/10.1080/17461391.2014.982204

Published online: 02 Dec 2014.

Submit your article to this journal

Article views: 4359

View related articles

View Crossmark data

Citing articles: 26 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=tejs20
European Journal of Sport Science, 2016
Vol. 16, No. 1, 96–105, http://dx.doi.org/10.1080/17461391.2014.982204

ORIGINAL ARTICLE

Practice and play in the development of German top-level professional

football players

MANUEL HORNIG1,2,3, FRIEDHELM AUST1, & ARNE GÜLLICH3

1
Institute of Sports Sciences, University of Mainz, Mainz, Germany, 2Department of Professional Football, Deutscher
Sportclub Arminia Bielefeld, Bielefeld, Germany, 3Department of Sport Science, University of Kaiserslautern, Kaiserslautern,
Germany

Abstract
This study examined the developmental sporting activities of 52 German football first Bundesliga professionals (including 18
senior national team members) and 50 fourth to sixth league amateur players. They reported their volumes of organised
football practice/training, including its “microstructure” (proportions of physical conditioning, skill exercises and playing
forms), non-organised leisure football play and engagement in other sports through their career, respectively. Analyses revealed
that the Bundesliga professionals performed moderate amounts of organised football practice/training throughout their career.
They accumulated 4264 (mean value) hours over ~16 years before debuting in 1st Bundesliga; senior National Team debut
was preceded by 4532 hours (mean) over ~17 years. Within the “microstructure” of organised practice/training, the proportion
of playing forms developed from ~52% (childhood) to ~45% (adolescence) and ~40% (adulthood) and physical conditioning
from ~13% to ~14% and ~23%. Outside organised involvement, these players engaged in extensive non-organised leisure
football play making ~68%, ~54% and ~9% of all football involvement. Subsuming organised and non-organised football,
~86% (childhood), ~73% (adolescence) and ~43% (adulthood) of all activity was game play (exclusive matchplay). National
Team differed from amateurs in more non-organised leisure football in childhood, more engagement in other sports in
adolescence, later specialisation, and in more organised football only at age 22+ years. Relative to numerous other studies,
these players performed less organised practice, particularly less physical conditioning, but greater proportions of playing
activities. The findings are discussed relative to the significance of playing forms and variable involvements and are reflected
against the deliberate practice and Developmental Model of Sport Participation (DMSP) frameworks.

Keywords: Professional football, deliberate practice, play, variability, individuality, talent development

Introduction models most frequently referenced (Bruner, Erick-

The role of different types and amounts of sporting
activities in developing elite performance is an issue of
debate in the literature. Namely, the significance of
sport-specific involvement and variable participation
in diverse sports performed in organised practice/
training or peer-led sporting leisure play (“deliberate
play”), respectively, are discussed (e.g. Côté, Lidor, &
Hackfort, 2009, for a review). Scholars have des‐
cribed the “macrostructure” of pathways characteris-
ing the development of these different activities
through the entire career, with “Deliberate Practice”
(DP; Ericsson, Krampe, & Tesch-Römer, 1993) and
the “Developmental Model of Sport Participation”
(DMSP; Côté, Baker, & Abernethy, 2007) being the
son, Wilson, & Côté, 2010).
Ericsson and colleagues suggested that essentially
DP designed and supervised by a coach, and
explicitly aiming to improve specific aspects of the
athlete’s performance through frequent repetition
and correction was the most effective type of activity
in acquiring high performance. They postulated a
minimum of 10,000 hours of DP over 10 years was
prerequisite to become an expert in any domain.
Moreover, they asserted the higher the performance
level, the greater the required amount of aggregated
DP (“monotonic benefits assumption”). Therefore,
aiming at the maximisation of DP aggregated
through the career implies extensive DP already in
childhood and its subsequent continuous expansion

Correspondence: A. Güllich, Technische Universität Kaiserslautern, Fachbereich Sozialwissenschaften, Fachgebiet Sportwissenschaft,

Erwin-Schrödinger-Straße Geb. 57, D-67663 Kaiserslautern, Germany. E-mail: guellich@sowi.uni-kl.de

© 2014 European College of Sport Science

 Development of German top-level professional football players
(p. 368, 387, 388; in sports: see Baker & Young,
2014, for a review).
DMSP contrasts two distinct pathway models
leading to elite performance. The early “specialisa-
tion” pathway corresponds to the DP concept: early
concentration in one sport already within childhood
with reinforced sport-specific DP/training that is
subsequently expanded through adolescence and
adulthood. Alternatively, the “diversification” path-
way describes progression through three stages
(“sampling” 6–12, “specialising” 13–15, “invest-
ment” 16+ years) from initial involvement in mul-
tiple sports, little DP and large amounts of deliberate
play towards eventually specialising in one sport,
large amounts of sport-specific DP/training and little
or no deliberate play. Côté and colleagues (2007)
argued that diversified involvement may provide a
broad repertoire of physiological capacities, technical
and perceptual-tactical skills transferable across
related sports. Deliberate play is expected to amplify
variable playing experience and is also thought to
maximise joy and support sustainable motivation.
In addition, some authors pointed to the significance
of complementing the “macrostructure” by also con-
sidering the “microstructure” of different activity types
within organised sport-specific practice/training (e.g.
Deakin & Cobley, 2003; Williams & Hodges, 2005).
Problem and state of research
Both DP and DMSP build on (time) volume of
activity types during different age periods. Exploring
whether the models’ scope extends to contributing
to the explanation of success differences within elite
sport implies comparing the correspondence of more
and less successful elite athletes’ activity patterns to
either of the pathways. However, the strict definition
of DP was suggested earlier (Abernethy, Farrow, &
Berry, 2003) to be insufficient for application in the
sport domain and to require supplementation by
other activity types (e.g. play). In DMSP, the high
abstraction level and therefore relatively fuzzy
description of the pathways impede their direct
empirical testing. The empirical approach is therefore
rather to record well-defined activity types, compare
more and less successful performers, and subse-
quently reflect their correspondence to the pathway
models.
Researchers have usually used elite athletes’ retro-
spective recall to study their participation history.
There is, however, still considerable inconsistency in
the state of investigation from the last ~25 years. The
absolute practice/training volumes vary substantially
between different types of sports (e.g. artistic com-
position or endurance sports; Malina et al., 2013;
Seiler & Tonnesen, 2009; vs. team game sports;
Baker, Côté, & Abernethy, 2003; Güllich, 2014b;
97
Moesch, Trier Hauge, Wikman, & Elber, 2013).
Also, the significance of activity types to success
apparently varies across age and success ranges.
Reviewing 19 studies and their own investiga-
tions, Güllich and Emrich (2014) revealed – across
multiple sports – that accumulated sport-specific
practice/training differentiated success among sub-
adult athletes and among senior athletes only across
moderate and/or heterogeneous success ranges but
not senior national vs. world class. However, world
class mostly exhibited more variable involvements in
diverse sports and later specialisation. The authors
interpreted that top athletes aim to train at the
margin of their individual stress-tolerability over
many years without exceeding it and therefore,
rather cost-reducing, risk-buffering investment pat-
terns may raise the probability of long-term sustain-
able athletic careers.
The findings signify that conditions for elite senior
success cannot be concluded by simply extrapolating
the scope of findings from lower age and/or success
ranges. However, examinations of senior top perfor-
mers’ participation histories are scarce but are
required to investigate the acquisition of the highest
performance levels.
This paper explores developmental activities of
professional senior top football players. Despite
extensive previous investigation in elite football the
situation of research is still constrained and findings
are – just as in other sports – inconsistent. Most
notably, most studies were limited to youth players
(e.g. Ford et al., 2012; Ford, Ward, Hodges, &
Williams, 2009; Ford & Williams 2011; Ford, Yates,
& Williams, 2010; Partington & Cushion, 2013;
Ward, Hodges, Starkes, & Williams, 2007; Yates &
Williams, 2009). As far as senior players were involved
these studies considered activities only within selected
age ranges or activity types (Helsen, Starkes, &
Hodges, 1998; Zibung & Conzelmann, 2013) and/or
didn’t compare different success levels (Ford &
Williams, 2008; Leite, Baker, & Sampaio, 2009;
Memmert, Baker, & Bertsch, 2010). At a descriptive
level, previous research can be summarised as follows:
Most elite footballers started playing in early child-
hood. They typically performed moderate volumes of
organised football practice/training in childhood
(mean values roughly 100–300 hours × year–1 until
~13/14 years) and adolescence (~14–18 years ~250–
430 hours × year–1). This included ~35–45% (child-
hood) and ~25–40% (adolescence) playing forms.
Additional non-organised leisure football made a
volume of ~80–260% (childhood) and ~0–70% (ado-
lescence) and involvement in other sports ~20–80%
and ~20–60% of organised football.
Still, comparative examination of senior top
players’ embracive practice histories are lacking up
to date. The present study addresses this gap by
98 M. Hornig et al.
involving German 1st Bundesliga (including Na‐
tional Team) and senior amateur footballers. It
recorded comprehensive reports of the “macrostruc-
ture” of their developmental sporting activities sup-
plemented by the “microstructure” within organised
football practice/training. In light of the absence of
such studies in the literature, the nature of this
investigation is foremost explorative and no directed
a priori hypotheses are formulated, respectively. Yet
it allows for examining commonalities and differ-
ences between more and less successful players and
for subsequently relating the results to the DP and
DMSP pathway models.
Methods
The participants responded to a standardised ques-
tionnaire administered by mail. Potential partici-
pants were instructed that their response was
voluntary, all responses would remain anonymous,
and the use of the data was exclusively for scientific
purposes. This procedure received ethical approval
from the German Federal Institute of Sport Science.
Participants
In winter/spring of 2011, 50 players from three
fourth to sixth league teams (24.8 ± 4.1 years
[Mean ± SD] at that time) participated in this
survey. These are the highest amateur (AM) leagues
comprising the top ~2% of all German senior
amateur players. We intended to compare 1st
Bundesliga professionals to them. After our requests
to all 1st Bundesliga clubs had resulted in either no
reply or responses putting us off with vague pro-
spects for future cooperation (see also Meyer, 2006)
we elected to contact players personally and conduct
a “snowball” procedure. The first author has played
in various Bundesliga clubs. He contacted the
players he was personally acquainted with, mailed
them the questionnaire and also asked them to pass
it to other players. We are aware of potential short-
comings of this practice: the total number of
contacted players remains unknown and the sample
of volunteering players might be specifically select-
ive. Still we considered this approach justified in
view of lacking published participation histories of
current senior top footballers. In addition we
attended to the sample’s representativeness by com-
paring selected aspects to the complete population of
current German 1st Bundesliga players (see below).
This request resulted in 52 1st Bundesliga respon-
dents in summer/fall 2013 (1stBL; six goalkeepers,
30 defensive and 16 offensive outfielders): 29 cur-
rent and 23 former 1stBL players. The latter exited
1st Bundesliga during the last five (2.8 ± 1.4) years;
12 are still playing in different professional leagues,
11 have retired completely.
The 1stBL players were aged 30.7 ± 4.2 years at
the time of response. They were born and entered a
football club ~3 years before AM (1982.1 ± 4.2 vs.
1985.5 ± 4.2 and 1988.1 ± 4.2 vs. 1991.1 ± 4.1) and
first started playing in a regional (or higher) junior
league ~2 years earlier (1997.9 ± 4.2 vs. 2000.0 ±
3.8). They are, thus, still considered as part of one
player generation. In addition, 1stBL respondents
corresponded closely to the complete current 1st
Bundesliga population (Güllich, 2014a) in the age
for starting organised football practice (respondents
6.0 ± 2.0; population 5.8 ± 1.9 years), their age
of debut in a 1st Bundesliga team (21.8 ± 3.1 vs.
21.4 ± 3.0 years) and their highest transfer fee value
(60% <3 million, 23% 3–6, 17% >€6 million vs.
68%/16%/16%).
Some 18 respondents were members of the Ger-
man senior National Team (NT; FIFA-ranked 2nd–
5th in the respective period). They had also partici-
pated in UEFA Champions League and/or Europa
League and eight were Champions League finalists;
17 were champions or vice-champions in national
leagues/cups in Germany, Spain and/or England.
Measurement of career “milestones” and sporting
activities
We used a questionnaire described in earlier papers
(Güllich, 2014b; Güllich & Emrich, 2014) and
supplemented questions addressing the “microstruc-
ture” within organised football practice/training in
the 2013 Bundesliga survey. The respondents
reported the following variables:
. Age for start of organised football practice/
training in a football club, league/tournament
participation, and first nomination for a
regional and/or national representative team
. Frequency and duration of organised football
practice/training sessions in a club within the
age categories of ≤10, 11–14, 15–18, 19–21,
and 22+ years
. Proportions within organised football practice/
training time spent on (a) physical condition-
ing exercises (strength, power, speed, endur-
ance, flexibility), (b) instructed drill-like
exercises practising skills of ball control, pas-
sing, shooting, tackling, and (unidirectional)
re-enacting of extracted game sequences, and
(c) uninterrupted game play in terms of free
play, small-sided games, and conditioned
games (modified rules, pitch area, goals; e.g.
possession games, four goals, etc.) within the
 Development of German top-level professional football players
age categories of U11, U15, U19, and 22+
years
. Involvement in organised club-based practice/
training in other sports; non-organised sport-
ing leisure play outside sport clubs in football
and in other sports; in each case age periods of
involvement and frequency of the type of
activity within each age category, respectively
The instrument’s test-retest-reliability was examined
earlier (Güllich & Emrich, 2014) in a three-year
panel design involving 244 national squad athletes
from all Olympic sports (0.80≤rtt≤1.00). Twenty-
nine respondents’ reported training volumes were
also compared to the logs in their daily training
diaries over an entire one-year season (0.82≤r≤0.84).
In addition, current specific examination in football
involved 38 young players’ test-retest over three
weeks (16–23 years; 0.91≤rtt≤1.00).
Statistical analysis
Analyses were performed using SPSS 21.0.
Descriptive data include frequencies, mean value,
standard deviation, and Pearson’s V. The estimates
of the volumes of activities are reported as annually
accumulated time (organised football practice/train-
ing) or frequency (all activity types) by multiplying
the mean weekly time or frequency by the annual
weeks of involvement (Güllich, 2014b).
Group differences between each of the success
groups (NT, n = 18; remaining 1stBL, n = 34; AM,
n = 50) were analysed using chi-square (χ2) and
unpaired t-tests. A non-parametric U-test was con-
ducted for non-uniform (skewed) data distribution.
Proportions of activities within the “microstructure”
of football practice/training were compared across
age periods by paired t-tests.
Effect sizes are expressed as Cohen’s d using
pooled variance for differences in group means. All
statistical testing was two-tailed. A value of p < 0.05
was considered statistically significant.
Results
We first describe the 52 Bundesliga players’ involve-
ment and then compare the three success groups.
Description of the Bundesliga players
Some 51 respondents reported involvement in foot-
ball leisure play and 25 engaged in 1.8 ± 0.9 other
sports (21 organised, 19 non-organised). Of these,
15 engaged only in other game sports, three only
non-games, and seven in both. Other games
included invasion type (n = 8), net/wall (n = 20),
99
and target games (n = 3; sum >22 because nine
players participated in various game types).
The age structure of the career is presented in
Table I. All players began involvement in football,
both organised and non-organised, at ≤12 years.
They attained their first nomination for a represent-
ative junior team after ~8 years of organised football
practice/training. The 1stBL debut was preceded by
as long as ~16 years and the debut in senior NT by
>17 years of organised football involvement.
Most players (n = 41) continued football leisure
play until adolescence or adulthood. Engagement in
other sports also mostly started at ≤12 years, but
later than football. Most of these players (n = 19)
continued participation in various sports until late
youth or adulthood.
The absolute amount of organised football prac-
tice/training increased continuously while all other
activities decreased gradually from childhood
through youth and adulthood (Figure 1).
The Bundesliga players performed moderate
amounts of organised football practice/training
throughout their career, ranging from 104 ± 56
hours × year–1 until 10 years to 546 ± 125 hours ×
year–1 at 22+ years. They made their 1stBL debut
after accumulating 4264 ± 1631 hours of organised
sport-specific practice/training, and NT debut was
preceded by 4532 ± 1587 hours.
Interestingly, the players participated more fre-
quently in non-organised leisure football play than
organised practice/training until age 10 (186 ± 88 vs.
87 ± 37) and 11–14 years (157 ± 94 vs. 132 ± 51
times × year–1).
The proportions of physical conditioning, skill
exercising, and game play within organised in-club
practice/training sessions are presented in Figure 2.
Initially, 50–53% of the time was playing activity,
35–37% skill exercise, and 13% physical condition-
ing, and there was little differentiation between
preparation period and season. With advancing age,
the proportions of physical conditioning increased
and play decreased. The differentiation between
preparation period and season also became more
pronounced in terms of less physical conditioning
and more play during the season (3.34 < t < 8.53;
p < 0.01, respectively). Consequently, the increase of
physical conditioning along progressing age was
stronger in the preparation period (from 13% in
U11 to 37% at 22+ years; t = 8.94; p < 0.01) than in
the season (13–17%), and so was the decrease of
play (preparation period 50–32%; t = 5.81; season
53–42%; t = 3.45; p < 0.01). Still, forms of game
play remained the activity taking up most time
throughout the year until U19 and during the season
through all age ranges.
100 M. Hornig et al.

Table I. Age-related “milestones” of the sporting career of senior National Team (NT), further 1st Bundesliga (1stBL) and Amateur (AM,
fourth to sixth league) football players

NT + 1stBL NT 1stBL AM

Age at … [years]
Start football involvement 4.3 (1.1) 4.1 (0.5) 4.4 (1.3) 4.4 (1.1)
Start organised football practice/training 6.0 (2.0) 5.3 (1.1) 6.4 (2.2) 5.6 (2.0)
Start playing in a league/tournament 7.4 (2.8) 6.6 (1.1) 7.8 (3.2) 7.5 (2.5)
Start non-organised leisure football play 4.5 (1.5) 4.3 (1.0) 4.6 (1.7) 5.2 (2.4)
Cessation non-organised leisure football play 18.1 (4.0) 17.7 (5.0) 18.3 (3.5) 18.5 (4.1)
First nomination for a regional or national representative junior teama 14.1 (2.3) 14.0 (2.1) 14.2 (2.4) 16.6 (2.1)
Debut in 1st Bundesligab 21.8 (3.1) 20.4 (3.3) 22.6 (2.8)
Debut in senior National Team 22.7 (2.3)
Start other sports 7.8 (2.6) 8.6 (2.1) 7.2 (2.8) 9.7 (5.4)
Cessation other sports 18.1 (6.1) 19.9 (6.1) 16.9 (5.9) 18.4 (5.7)
Start other sports organised practice/training 8.2 (2.0) 8.7 (1.7) 7.8 (2.3) 9.4 (4.4)
Cessation other sports organised 16.3 (6.1) 18.1 (6.3) 14.8 (5.8) 16.7 (5.5)
Start other sports non-organised leisure play 8.7 (3.5) 10.6 (3.3) 7.1 (2.9) 10.6 (5.5)
Cessation other sports non-organised play 17.2 (4.8) 17.5 (5.4) 16.9 (4.6) 18.1 (5.8)
Specialise exclusively in footballc 12.1 (7.6) 14.3 (9.2) 10.9 (6.5) 9.9 (7.2)
Duration of involvement in organised football until … [years]
First nomination for a regional or national representative junior teamd 8.2 (2.7) 8.8 (2.7) 7.9 (2.7) 11.3 (2.6)
Debut in 1st Bundesliga 15.9 (3.0) 15.1 (3.4) 16.3 (2.8)
Debut in senior National Team 17.4 (2.5)
Specialise exclusively in footballe 6.2 (7.8) 9.2 (8.9) 4.6 (6.8) 4.1 (6.9)
a
NT vs. AM: t = 4.01; p < 0.01; d = 1.24; 1stBL vs. AM: t = 3.48; p < 0.01; d = 1.06.
b
NT vs. 1stBL: t = 2.45; p < 0.05; d = 0.70.
c
NT vs. AM: t = 2.08; p < 0.05; d = 0.54.
d
NT vs. AM: t = 3.08; p < 0.01; d = 0.95; 1stBL vs. AM: t = 4.79; p < 0.01; d = 1.26.
e
NT vs. AM: Z = 2.19; p < 0.05; d = 0.64.
Mean (±SD) values. Age for start and cessation of leisure football play and involvement in other sports is calculated only for the players
involved in the respective activity; duration figures are based on all respondents including those with a value of 0; likewise, age of
specialisation is calculated for all players including those without involvement in other sports.

Subsuming the absolute frequency of non-
organised football play and the proportion of
game play within organised sessions, an estimate of
roughly 86% ± 12% (–10 years), ~73% ± 13% (–14),
~58% ± 14% (–18) and ~43% ± 14% (22+ years) of
all football involvement was playing activity (exclus-
ive match play, respectively). Within all football
play, the share performed within in-club sessions
developed from ~20% ± 19% (–10 years) to ~34% ±
22% (–14), ~63% ± 28% (–18) and ~91% ± 21%
(22+ years).
Comparison between success levels
No significant differences were revealed between the
success groups in the numbers of players engaging in
the respective types of activities or sports, the
“microstructure” within organised football or the
total volume of all sport involvement or football
involvement (organised plus non-organised, respect-
ively) at any age.
NT and 1stBL players differed from AM,
respectively, in significantly earlier nominations
for junior representative teams and NT debuted
earlier in 1st Bundesliga than 1stBL (Table I). NT
specialised exclusively in football after ~9 years of
involvement which was significantly later than their
peers.
NT and 1stBL players participated in significantly
more non-organised leisure football play until age 10
years than AM, respectively (Figure 3). 1stBL
exhibited a stronger increase of organised football
practice/training during adolescence/adulthood than
AM and NT. Interestingly, NT did not differ
significantly from AM in this regard before 22+
years; but NT players performed significantly more
organised practice/training sessions in other sports
through adolescence compared to their AM and also
1stBL peers.
It is also important to note that these attributes
varied considerably within each success group,
respectively. The duration of organised football
involvement before a nomination for a junior rep-
resentative team ranged from 5 to 16 years within
NT (V = 31%), 3–15 among 1stBL (V = 34%), and
5–15 years in AM (V = 21%); and before debuting
in 1st Bundesliga 11–19 (NT; V = 22%) and 10–25
years (1stBL; V = 17%). The volume of organised
 Development of German top-level professional football players 101

600 (A) FB org me

Hours p.a.
400

200

400 (B) FB org freq (D) OS org freq

Frequency p.a.

300 40

200
20
100

0 0

(C) FB leis freq 80 (E) OS leis freq

Frequency p.a.

200 60

40
100
20

0 0
-10 11-14 15-18 19-21 22+ -10 11-14 15-18 19-21 22+
Age Category [years] Age Category [years]

Figure 1. Developmental activities of 1st Bundesliga football players (including senior National Team members) within defined age categories.
Absolute volume (time, frequency per year) of organised in-club football practice/training (A, B: FB org), non-organised football leisure play
(C: FB leis), organised in-club involvement in other sports (D: OS org) and non-organised leisure activity in other sports (E: OS leis).
Figures represent mean values and standard deviation (error bars). Note the different ordinate scales in (A) to (E).

Physical Condioning Skill Exercises Game Play

Preparaon Period Season
100%

32%
Percentage of Pracce Time

80% 37% 42%

43% 46% 47%
50% 53%

60%
31%
33%
40% 41%
40%
42% 39%
37% 35%
20% 37%
30%
13% 17% 13% 13% 15% 17%
0%
U11 U15 U19 22+ U11 U15 U19 22+
Age Category

Figure 2. Proportions of physical conditioning, skill exercises, and game play within organised in-club football practice/training sessions.
Mean values of 1st Bundesliga football players (including senior National Team members). Percentage is calculated exclusive time for
warm-up activities, respectively.

football accumulated until 21 years ranged from non-organised football until age 21 was 720–5348
2594 to 5052 hours (NT; V = 19%), 2044–7920 (NT; V = 49%), 0–8448 (1stBL; V = 69%) and
(1stBL; V = 31%) and 1767–5027 hours (AM; V = 0–7820 (AM; V = 73%), and of participation in
22%). The range of the frequency of participation in other sports 0–2420 (NT; V = 122%), 0–1680
102 M. Hornig et al.
NT 1st BL AM Group comparisons

600 (A) FB org me **

## ## † NT vs. 1 BL
st
* – p<0.05, ** – p<0.01
NT vs. AM † – p<0.05

Hours p.a.
# st
1 BL vs. AM # – p<0.05, ## – p<0.01
400

200

400 (B) FB org freq **

##
##
† † * (D) OS org freq
30
Frequency p.a.

300

20
200

100 10 *

0 0

#† (C) FB leis freq (E) OS leis freq

200
30
Frequency p.a.

20
100
10

0 0
-10 11-14 15-18 19-21 22+ -10 11-14 15-18 19-21 22+
Age Category [years] Age Category [years]

Figure 3. Absolute volume (time, frequency per year) of organised in-club football practice/training (A, B: FB org), non-organised football
leisure play (C: FB leis), organised in-club involvement in other sports (D: OS org) and non-organised leisure activity in other sports (E: OS
leis) among senior National Team (NT), other 1st Bundesliga (1stBL) and Amateur (AM, fourth to sixth league) football players within
defined age categories.
Figures represent mean values (standard deviation omitted for clarity). Note the different ordinate scales in (A) to (E). Organised football
time: NT vs. 1stBL 19–21 years t = 2.87; p < 0.01; d = 0.88. NT vs. AM 22+ years t = 2.44; p < 0.05; d = 0.75. 1stBL vs. AM
15–18 years Z = 1.99; p < 0.05; d = 0.54; 19–21 years t = 4.27; p < 0.01; d = 1.01; 22+ years t = 4.18; p < 0.01; d = 1.07. Organised football
frequency: NT vs. AM 19–21 years t = 2.78; p < 0.01; d = 0.86. 1stBL vs. AM 19–21 years t = 4.23; p < 0.01; d = 1.00; 22+ years t = 4.15;
p < 0.01; d = 1.06. Non-organised leisure football: NT vs. AM –10 years Z = 2.20; p < 0.05; d = 0.58. 1stBL vs. AM –10 years t = 2.04; p <
0.05; d = 0.46. Organised other sports: NT vs. 1stBL 15–18 years Z = 2.15; p < 0.05; d = 0.57; 19–21 years Z = 2.50; p < 0.05; d = 0.64.
NT vs. AM L 11–14 years Z = 2.02; p < 0.05; d = 0.53; 15–18 years Z = 2.25; p < 0.05; d = 0.60.

(1stBL; V = 156%) and 0–3764 (AM; V = 205%),
respectively.
Discussion
The central results of this study are that the senior
top-level footballers started involvement in early
childhood and performed moderate volumes of
organised football practice/training throughout their
career, with extensive game play in terms of large
proportions of playing forms within organised in-
club practice/training through all age periods,
substantial non-organised leisure football play dur-
ing childhood and youth and in many cases also
participation in game sports other than football.
Overall, the success groups’ participation histories
displayed much commonality. Significant differences
existed in that National Team specialised later and
performed more leisure football in childhood, more
other sports in adolescence, and more organised
football only in adulthood compared to Amateurs.
First Bundesliga differed from Amateurs in more
leisure football in childhood and more organised
football in adolescence and adulthood. However,
 Development of German top-level professional football players
each type of activity varied substantially within each
success group.
From an empirical perspective, the absolute
volumes of football practice/training are consistent
with the observation that elite players in team game
sports typically perform less practice/training than
those in some other types of (individual) sports (see
introduction) but engage extensively in sporting
leisure play. The present findings also correspond
to other studies in various sports showing that world
class athletes mostly did not differ from national
class in more sport-specific practice/training, but in
more variable involvements (Carlson, 1988; Güllich,
2014b; Güllich & Emrich, 2014; Johnson, 2006;
VanRossum, 2000).
Relative to other elite football samples, the present
players’ absolute volumes of organised football prac-
tice/training and their proportion of non-organised
leisure football corresponded to those of other
German and Swiss senior elite players (Gonzalez-
Balzar, 2007; Memmert et al., 2010; Zibung &
Conzelmann, 2013) within the respective age ranges.
However, they performed significantly less sport-
specific in-club practice/training and greater propor-
tions of non-organised play than numerous other
samples (Ford et al., 2009, 2012; Ford & Williams
2011; Helsen et al., 1998; Huijgen, Elfering-
Gemser, Lemmink, & Vischer, 2014; Huijgen, Elfer-
ing-Gemser, Post, & Vischer, 2009, 2010; Ward
et al., 2007). These differences are plausible as the
referenced samples were youth elite footballers
(except Helsen et al., 1998) who joined youth
academies or equivalent institutions already before
age 15, which applied to only 12% of the present
Bundesliga players (consistent in Premier League:
Anderson & Miller, 2011). Also, within the “micro-
structure” of organised practice/training, the respon-
dents’ proportion of playing forms ranged at the
upper margin or above other reports while phys-
ical conditioning was far below other samples during
childhood/adolescence (Ford et al., 2010; Partington &
Cushion, 2013; Ward et al., 2007; Yates &
Williams, 2009).
From a theoretical perspective, the matter of inter-
est is favourable compositions and proportions of
activity types along progressing age rather than any
single activity’s significance. The present top footbal-
lers mostly combined moderate volumes of
instructed exercise with extensive forms of game
play, whereas play shifted in large parts from non-
organised play and other game sports in childhood/
youth towards organised in-club play in adulthood.
Unlike exercise, play involves the interaction
between motor, cognitive-perceptual and decision-
making skills and their variable and random execu-
tion in a variety of similar but different action
sequences and situational contexts and thereby
103
also provides extensive implicit skill learning (e.g.
Davids, Button, & Bennett, 2008; Williams &
Hodges, 2005). These characteristics facilitate the
development of adaptable and transferable skills that
can be applied in varying situational contexts. They
are also more robust in terms of less susceptibility to
detrimental effects of physiological or psychological
stress and more sustainable in terms of better
retention and of promoting subsequent contin-
ued learning and skill refinement, compared to
skills acquired through more standardised drill-like
exercise (Schmidt & Lee, 1999; Williams & Hodges,
2005). Uninstructed sporting leisure play, for its
part, is assumed to add further amplification to
participants’ variability of playing experience. The
dimensions of the field and goals and the number of
players vary, the participants are often heterogeneous
in age and skill, and the players typically experience
playing on different positions, different grounds, and
with different balls. It may also facilitate discovery
learning in terms of providing opportunities to
experiment varying technical and tactical variants
and find out individual solutions through trial and
error (Davids et al., 2008).
Finally, from a conceptual perspective, first, the only
activity type (in parts) corresponding to the original
definition of DP – organised sport-specific practice/
training – accumulated to only ~4500 hours – thereof
~45% play, i.e. non-DP. In addition, respective
differences between success groups existed only in
later age categories and their direction is clearly at
odds with the “monotonic benefits assumption”
(1stBL>NT), while there were significant success-
related differences in non-DP activities. Second,
most players’ developmental participation patterns
corresponded only partly to and partly deviated from
the attributes of either pathway model described in
DMSP. Also, most footballers continued engage-
ment in leisure football and in other sports signifi-
cantly longer than predicted by DMSP, and their
development was characterised by continuous gradual
increase or decrease of the respective activities rather
than reflecting transition points between distinguish-
able stages. Third, sizeable variation of each activity
type indicated great inter-individual differences
within each success group. All combinations of
more or less engagement in organised football, non-
organised football and other sports were existent
among more and less successful footballers, respect-
ively. That is, a multitude of different “paths” may
lead to top success while more or less close corres-
pondence to any of the pathways described in DP and
DMSP provides little or no explanatory power to
explain success differences in German elite football at
the individual athlete level.
104 M. Hornig et al.
Methodological considerations and future directions
While the retrospective approach is the method of
choice to encompass respondents’ entire career,
some limitations inherent to the approach and to
the subject should be considered. E.g.:
. Involving current top senior players, potential
recent changes in juvenile conditions for future
senior success cannot be captured.
. Also, potential earlier dropouts are not included.
. Essentially, the cross-sectional design impedes
causal interpretation.
Furthermore, the findings of this and numerous
other studies (e.g. reviews in Baker & Young, 2014;
Güllich & Emrich, 2014) point to the limitations of
mere quantities of relatively coarse categories of
activity types and of group comparisons within the
top ranges of elite sport as a field of profound
individuality. Subsequent research may pursue the
question which attributes constitute the individuality
that moderates responsiveness to different config-
urations of activities in order to better understand
why which composition of which amounts of which
activities are beneficial to which athlete during which
age period. Clearly, beyond physical/physiological
attributes this requires drawing in the athlete’s
interaction with his/her social environment. For
example, in light of the significance of the concept
of long-term sustainability hypothesised by Güllich
and Emrich (2014), continuous balance of physical
and psychological strain with individual stress toler-
ability as well as balancing time expenditure in sports
with time demands external to sport (education,
family, friends, hobbies, etc.) over multiple years
appear to warrant particular consideration. This
implies documenting performed activities with
higher temporal resolution and more “fine-grained”,
i.e. dimensionally rather than just categorically
(dimensions for example degree of relatedness to
the demands of the competition, degree of variabil-
ity, physical and mental demand, inherent enjoy-
ment, and degree of being instructed and organised),
while also taking qualitative attributes into account.
Acknowledgement
We would like to express our sincere thanks to the
anonymous reviewers for fruitful comments on
an earlier version of this paper. We would also
like to acknowledge and thank Martin Andes,
Florian Cordier, Franziska Frank, Sajoscha Hir-
mann, Tamara Larkins, Paul Messer, Sören Schan-
zenbecher, Christoph Schreieck, and Johannes
Spitz, master students in physical education at the
University of Kaiserslautern and members of the
project team, for their efforts in the collection and
preparation of the data.
References
Abernethy, B., Farrow, D., & Berry, J. (2003). Constraints and
issues in the development of a general theory of expert
perceptual-motor performance: A critique of the deliberate
practice framework. In J. L. Starkes & K. A. Ericsson (Eds.),
Expert performance in sports: Advances in research on sport expertise
(pp. 349–369). Champaign, IL: Human Kinetics.
Anderson, G., & Miller, R. M. (2011). The academy system in English
professional football: Business value or “Following the Herd”? research
paper series no. 2011/43. Liverpool: University of Liverpool
Management School. Retrieved from www.liv.ac.uk/manage-
mentschool/research/working%2520papers/wp201143.pdf
Baker, J., Cote, J., & Abernethy, B. (2003). Sport-specific
practice and the development of expert decision-making in
team ball sports. Journal of Applied Sport Psychology, 15(1),
12–25. doi:10.1080/10413200305400
Baker, J., & Young, B. (2014). 20 years later: Deliberate practice
and the development of expertise in sport. International Review
of Sport and Exercise Psychology, 7(1), 135–157. doi:10.1080/
1750984X.2014.896024
Bruner, M. W., Erickson, K., Wilson, B., & Côté, J. (2010). An
appraisal of athlete development models through citation
network analysis. Psychology of Sport and Exercise, 11(2), 133–
139. doi:10.1016/j.psychsport.2009.05.008
Carlson, R. C. (1988). The socialization of elite tennis players in
Sweden: An analysis of the players’ background and develop-
ment. Sociology of Sport Journal, 5, 241–256.
Côté, J., Baker, J., & Abernethy, B. (2007). Practice and play in
the development of sport expertise. In R. Eklund &
G. Tenenbaum (Eds.), Handbook of sport psychology (3rd ed.,
pp.184–202). Hoboken, NJ: Wiley.
Côté, J., Lidor, R., & Hackfort, D. (2009). ISSP position stand:
To sample or to specialize? Seven postulates about youth sport
activities that lead to continued participation and elite perform-
ance. International Journal of Sport and Exercise Psychology, 9,
7–17. doi:10.1080/1612197X.2009.9671889
Davids, K. W., Button, C., & Bennett, S. J. (2008). Dynamics of
skill acquisition: A constraints-led approach. Champaign, IL:
Human Kinetics.
Deakin, J. M., & Cobley, S. (2003). A search for deliberate
practice. An examination of the practice environments in figure
skating and volleyball. In J. L. Starkes & K. A. Ericsson (Eds.),
Expert performance in sports: Advances in research on sport expertise
(pp.115–135). Champaign, IL: Human Kinetics.
Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). The
role of deliberate practice in the acquisition of expert perform-
ance. Psychological Review, 100, 363–406. doi:10.1037/0033-
295X.100.3.363
Ford, P. R., Carling, C., Garces, M., Marques, M., Migueal, C.,
Farrant, A., … Williams, A. M. (2012). The developmental
activities of elite soccer players aged under-16 years from
Brazil, England, France, Ghana, Mexico, Portugal and Swe-
den. Journal of Sports Sciences, 30, 1653–1663. doi:10.1080/
02640414.2012.701762
Ford, P. R., Ward, P., Hodges, N. J., & Williams, A. M. (2009).
The role of deliberate practice and play in career progression in
sport: The early engagement hypothesis. High Ability Studies,
20(1), 65–75. doi:10.1080/13598130902860721
Ford, P. R., & Williams, A. M. (2008). The effect of participa-
tion in Gaelic football on the development of Irish profes-
sional soccer players. Journal of Sport & Exercise Psychology,
30, 709–722.
 Development of German top-level professional football players
Ford, P. R., & Williams, A. M. (2011). The developmental
activities engaged in by elite youth soccer players who pro-
gressed to professional status compared to those who did not.
Psychology of Sport and Exercise, 13, 349–352. doi:10.1016/j.
psychsport.2011.09.004
Ford, P. R., Yates, I., & Williams, A. M. (2010). An analysis of
practice activities and instructional behaviours used by youth
soccer coaches during practice: Exploring the link between
science and application. Journal of Sports Sciences, 28, 483–495.
doi:10.1080/02640410903582750
Gonzalez-Balzar, P. (2007). Konditionstraining und sportmedizinische
Betreuung in der 1. Fußballbundesliga [Physical conditioning and
sports-medical support in football 1st Bundesliga]. Hamburg:
University of Hamburg (doctoral dissertation). Retrieved from
ediss.sub.uni-hamburg.de/frontdoor.php?source=3575
Güllich, A. (2014a). Selection, de-selection and progression in
German football talent promotion. European Journal of Sport
Science, 14, 530–537. doi:10.1080/17461391.2013.858371
Güllich, A. (2014b). Many roads lead to Rome – Developmental
paths to Olympic gold in men’s field hockey. European Journal of
Sport Science, 14, 763–771. doi:10.1080/17461391.2014.905983
Güllich, A., & Emrich, E. (2014). Considering long-term sustain-
ability in the development of world class success. European
Journal of Sport Science, 14 (Suppl. 1), S383–S397. doi:10.1080/
17461391.2012.706320
Helsen, W. F., Starkes, J. L., & Hodges, N. J. (1998). Team
sports and the theory of deliberate practice. Journal of Sport &
Exercise Psychology, 20, 12–34.
Huijgen, B. C. H., Elfering-Gemser, M., Lemmink, K. P. A. M.,
& Vischer, C. (2014). Multidimensional performance charac-
teristics in selected and deselected talented soccer players.
European Journal of Sport Science, 14, 2–10. doi:10.1080/1746
1391.2012.725102
Huijgen, B. C. H., Elferink-Gemser, M., Post, W., & Visscher, C.
(2009). Soccer skill development in professionals. International
Journal of Sports Medicine, 30, 585–591. doi:10.1055/s-0029-
1202354
Huijgen, B. C. H., Elfering-Gemser, M., Post, W. J., & Vischer,
C. (2010). Development of dribbling in talented youth soccer
players aged 12–19 years: A longitudinal study. Journal of Sports
Sciences, 28, 689–698. doi:10.1080/02640411003645679
Johnson, M. B. (2006). Climbing to the top: Uncovering processes that
facilitate and inhibit attaining excellence. Tallahassee: Florida
State University (doctoral dissertation). Retrieved from http://
diginole.lib.fsu.edu/etd
Leite, N., Baker, J., & Sampaio, J. (2009). Paths to expertise in
Portuguese national team athletes. Journal of Sport Science &
Medicine, 8, 560–566.
105
Malina, R. M., Baxter-Jones, A. D. G., Armstrong, N., Beunen,
G. P., Caine, D., Daly, R. M., … Russell, K. (2013). Role of
intensive training in the growth and maturation of artistic
gymnasts. Sports Medicine, 43, 783–802. doi:10.1007/s40279-
013-0058-5
Memmert, D., Baker, J., & Bertsch, C. (2010). Play and practice
in the development of sport-specific creativity in team ball
sports. High Ability Studies, 21(1), 3–18. doi:10.1080/13598
139.2010.488083
Meyer, T. (2006). Organizing elite soccer training – Scientific
knowledge vs. soccer-specific tradition [in German: Trainings-
gestaltung im Leistungsfußball – wissenschaftliche Erkennt-
nisse vs. sportartspezifische Tradition]. Deutsche Zeitschrift für
Sportmedizin, 57, 132–137.
Moesch, K., Trier Hauge, M.-L., Wikman, J. M., & Elber, A.
(2013). Making it to the top in team sports: Start later,
intensify, and be determinded! Talent Development and Excel-
lence, 5(2), 85–100.
Partington, M., & Cushion, C. (2013). An investigation of the
practice activities and coaching behaviors of professional top-
level youth soccer coaches practice activities and behaviors of
youth soccer coaches. Scandinavian Journal of Medicine &
Science in Sports, 23, 374–382. doi:10.1111/j.1600-0838.2011.
01383.x
Schmidt, R. A., & Lee, T. A. (1999). Motor control and learning: A
behavioral emphasis. Champaign, IL: Human Kinetics.
Seiler, S., & Tonnesen, E. (2009). Intervals, thresholds, and long
slow distance: The role of intensity and duration in endurance
training. Sportscience, 13, 32–53. Retrieved from sportsci.org/
2009/ss.htm
Van Rossum, J. H. A. (2000). Deliberate practice and Dutch field
hockey: An addendum to Starkes. International Journal of Sport
Psychology, 31, 452–460.
Ward, P., Hodges, N. J., Starkes, J. L., & Williams, A. M. (2007).
The road to excellence: Deliberate practice and the develop-
ment of expertise. High Ability Studies, 18, 119–153.
doi:10.1080/13598130701709715
Williams, A. M., & Hodges, N. J. (2005). Practice, instruction
and skill acquisition in soccer: Challenging tradition. Journal
of Sports Sciences, 23, 637–650. doi:10.1080/0264041040002
1328
Yates, I., & Williams, A. M. (2009). The microstructure of
practice in soccer. In T. Reilly & F. Korkusuz (Eds.), Science
and Football VI (pp.442–446). London: Routledge.
Zibung, M., & Conzelmann, A. (2013). The role of specialisation in
the promotion of young football talents: A person-oriented study.
European Journal of Sport Science, 13, 452–460. doi:10.1080/
17461391.2012.749947