You are on page 1of 19

ZU-TH 11/12 SLAC-PUB 15190 IPPP/12/51 DCPT/12/102 LPN12-080 FR-PHENO-2012-018 MCNET-12-08

arXiv:1207.5031v2 [hep-ph] 13 Feb 2013

NLO QCD matrix elements + parton showers in e+e− → hadrons
Thomas Gehrmann1 , Stefan H¨ oche2 , Frank Krauss3 , Marek Sch¨ onherr3 , Frank Siegert4
1 2 3 4

Institut f¨ ur Theoretische Physik, Universit¨ at Z¨ urich, CH-8057 Z¨ urich, Switzerland SLAC National Accelerator Laboratory, Menlo Park, CA 94025, USA Institute for Particle Physics Phenomenology, Durham University, Durham DH1 3LE, UK Physikalisches Institut, Albert-Ludwigs-Universit¨ at Freiburg, Hermann-Herder-Str. 3, D-79104 Freiburg, Germany

Abstract: We present a new approach to combine multiple NLO parton-level calculations
matched to parton showers into a single inclusive event sample. The method provides a description of hard multi-jet configurations at next-to leading order in the perturbative expansion of QCD, and it is supplemented with the all-orders resummed modelling of jet fragmentation provided by the parton shower. The formal accuracy of this technique is discussed in detail, invoking the example of electron-positron annihilation into hadrons. We focus on the effect of renormalisation scale variations in particular. Comparison with experimental data from LEP underlines that this novel formalism describes data with a theoretical accuracy that has hitherto not been achieved in standard Monte Carlo event generators.

1

Introduction

During the past decade, Monte-Carlo methods for simulating hadronic final states in collider experiments have improved continuously. Multi-purpose event generators incorporating the most recent higher-order perturbative QCD calculations have thus emerged, making them available to phenomenology and experiment alike. This has far-reaching consequences for both precision physics and searches for new phenomena. Key to the developments has been the steady progress in understanding the interplay of real and virtual higherorder QCD corrections on one hand and of resummation techniques like parton-shower algorithms on the other hand. The construction and development of simulation tools for QCD processes has become one of the central activities of research in collider phenomenology. This publication discusses an extension to the established techniques of multi-jet merging and next-toleading order matrix-element matching. Existing multi-jet merging methods (MEPS) combine leading-order matrix elements of varying final-state multiplicity with parton showers. They were pioneered in [1–4] and further matured in [5–8]. The key advantage of these methods is the possibility to describe arbitrarily

The outline of the present paper is as follows: Section 2 discusses the MEPS algorithm for matrix-element merging at leading order.2.j = pi pk . In the framework of this paper the formalism is specified for a multi-jet merging at NLO accuracy for e+ e− -annihilations into hadrons. This is remedied by next-to-leading order matrix-element matching methods (MC@NLO). 10] and have recently been automated in various programs [11. Our approach is justified by the choice of transverse momentum as evolution variable in the parton shower used for this publication. 4 is devoted to details of the Monte-Carlo implementation. equivalent to the measure of hardness of a parton splitting. which combine NLO QCD calculations of fixed jet multiplicity with parton showers. building on the existing implementations of MEPS [5] and MC@NLO [12] techniques in the SHERPA event generator [15. 12].j j. Qcut denotes a jet-defining cut value. As an intermediate step. in order to reflect the fact that no unique parton flavour can be assigned at the next-to-leading order. that effects due to a mismatch of these two quantities can be neglected. they lack the precision of a full next-to-leading order perturbative calculation. the new merging method at next-to leading order.i where k Ci. and the MC@NLO method for NLO matching as implemented in SHERPA. 3. which typically denotes the minimal value of some relative transverse momentum present in the n-parton phase-space configuration Φn . allowing to focus on the structure of the result. (pi + pk )pj (2. 5. 1 The jet criterion Q applied here has been slightly modified compared to [5]. including the impact of scale variations and of varying the number of jets described by NLO matrix element calculations. also for e+ e− -annihilations into hadrons.complex final states at leading order in the strong coupling. however. which improves the readability of this publication. 16]. The renormalisation scale dependence of the result is discussed in some detail. 6 presents our conclusions. 14].10] are ignored. For a full algorithmic solution. 2 . the quantity of interest is the expectation value O of an arbitrary. we will assume that the evolution parameter of the parton shower is defined in such a way. called the merging scale. Sec. In other words we will neglect effects that arise from allowed emissions generated by truncated parton showers. it is possible to make these results exclusive and combine them with higher-multiplicity leading-order predictions in order to recover the virtues of MEPS methods. The new method discussed here goes well beyond the scope of the MENLOPS technique. infrared-safe observable O. Sec.1. However. Correspondingly. processindependent merging method for matched NLO predictions with varying jet multiplicity. such that for n-jet events the condition Qn > Qcut is fulfilled1 . we intend to replace the leading-order matrix elements of the original MEPS approach with corresponding ones at next-to-leading order. is introduced in Sec. The aim of this paper is to further improve upon the existing algorithms and construct a consistent. An algorithm with the same goals and a similar setup for the parton shower has been detailed. existing matrix-element parton-shower merging and matching methods are briefly reviewed.1) The spectator index k runs over all possible coloured particles. inclusive final states. For any pair of final-state partons i and j we define Q2 ij = 2 pi pj min k=i. This is achieved by combining various MC@NLO event samples and accounting for potential double counting by means of a modified truncated parton shower [5. while [19] proposed a general method for NLO vector boson production plus n jets and implemented it for n=0. 10]. we intend to maintain the fixed-order accuracy of the matrix elements. the effects of allowed emissions generated by truncated showers [5. As already stated in the introduction. while providing fully inclusive event samples with resummation effects taken into account. Here we apply the method of [19] to hadronic final states in e+ e− -annihilation. evaluated by taking the average over sufficiently many points in an n-particle phase-space. MEPS@NLO. In the present paper. using the notation of [12].j 2 k + Ck Ci. Pictorially speaking. 2 Brief review of merging and matching techniques In this section. Ultimately. we refer to the parallel publication. They have therefore become standard analysis tools for collider experiments. They were introduced in [9. but also to preserve the logarithmic accuracy of the parton shower. A method for merging NLO vector boson plus 0 and 1-jet samples was introduced in [18]. Formally. we define a jet criterion Qn . Example results for the case of electron-positron annihilation into hadrons are shown in Sec. In the context of merging. With the notation thus established. Φn . Their main advantage lies in the excellent description of well-defined. the implementation of the MENLOPS idea for MC@NLO core processes is presented. in [19]. Using the MENLOPS technique [13. in [17].

µ2 n Q ) Θ(Qcut − Qn+1 ) O (Φn+1 ) dΦn+1 Bn+1 (Φn+1 ) ∆(PS) (tn+1 . Bn (Φn ). tc is an infrared regulator of the order of ΛQCD marking the transition into the non-perturbative region. is parametrised as dΦ1 = dt dz dφ J (t. µ2 n Q ) Θ(Qn+1 − Qcut ) O (Φn+1 ) . • Sudakov form factors of the parton shower.3) where t is the ordering variable. t n ) = exp − t dΦ1 Kn (Φ1 ) . • The resummation scale µQ . to O(αs ) has been computed in [14]. infrared-finite observable O. while for MEPS@NLO leading-order accuracy is envisaged. have the following aims • Multi-jet merging techniques For configurations with n jets. For next-to-leading order merging (MEPS@NLO) they should be given at NLO accuracy. summed (averaged) over final state (initial state) spins and colours and including symmetry and flux factors.1 Leading-order merging . The term on the second line yields leading-order accuracy for any n +1-particle observable in the region Qn+1 > Qcut . jet observables for n jets above the merging scale Qcut should be determined at leading-order accuracy. while the terms dΦn Bn and dΦn+1 Bn+1 correspond to the fixed-order event generation. φ) is 2 the appropriate Jacobian factor. leading order for n partons. as well as our newly proposed technique. the following quantities are introduced: • Squared leading-order (Born) matrix elements. • NLO matching methods For processes leading to n-parton final states at leading order all n-particle inclusive observables.4) where O(Φm ) is the observable evaluated for an m-parton final state. At the same time we require that the logarithmic accuracy of the shower be either maintained or improved in the region above Qcut . z. (2. The square bracket on the first line and the Sudakov factor on the second line are both generated by the parton shower. The expectation value of an arbitrary. It is derived from the following expression: µ2 Q O = + dΦn Bn (Φn ) ∆(PS) (tc . which defines an upper limit of parton evolution in terms of the shower evolution variable. More precisely. for n outgoing particles. for leading-order merging (MEPS).MEPS In the context of the leading-order merging method proposed in [5]. 2.2) Kn denotes the sum of all splitting kernels for the n-body final state. J (t.The methods reviewed here. The ordering variable is usually taken to fulfil t ∝ k⊥ as t → 0. µ2 n Q ) O (Φn ) + tc dΦ1 Kn (Φ1 ) ∆(PS) (tn+1 . At the same time. dΦ1 . z. the MEPS accuracy will be that of the parton shower. while higher-order emissions should still be described by the leading logarithmic approximation of the parton shower. (2. and in particular the total cross section. the respective fixed-order accuracy of O inherent to the parton-level result should be maintained. z is the splitting variable and φ is the azimuthal angle. all n +1-particle observables are expected to be given at leading order accuracy. φ) . given by t ∆(PS) (t. are expected to reproduce the fixed order NLO results. (2. Leading-order accuracy for observables 3 . The one-particle phase-space element for a splitting. For configurations below Qcut .

The size of these corrections determines the potential discontinuity in O at Qcut . where each dipole encodes exactly one singular region [12]. µ2 n Q ) O (Φn ) + tc dΦ1 Kn (Φ1 ) ∆(PS) (tn+1 . By replacing n → n + 1.6) + dΦ1 2 D(A) n (Φn+1 ) Θ(µQ − tn+1 ) − D(S) n (Φn+1 ) (S) . all properties of the algorithm remain the same. while Dn are the corresponding real subtraction terms. it is easily possible to obtain the exact same singularity structure as in the real-emission matrix element [21.3) in each of these dipole phase spaces. Φ1 ) B(Φn ) (A) . • The hard remainder function (A) 2 H(A) n (Φn+1 ) = Rn (Φn+1 ) − Dn (Φn+1 ) Θ(µQ − tn+1 ) . can be included in a systematic manner. Rn (Φn+1 ). (2. 2. µ2 n Q ) O (Φn+1 ) (2.4) can be rewritten as µ2 Q O = + dΦn Bn (Φn ) ∆(PS) (tc . Sub-leading colour configurations. each dipole has a corresponding phase space factorisation dΦn+1 = dΦn dΦij. (2. while the original Sudakov form factor ∆n does not. ¯ (A) • The NLO-weighted Born differential cross section B n . The additional terms on the second line incorporate possible sub-leading colour single logarithms as well as power corrections. which will be described in Sec. cf. • The MC@NLO Sudakov form factor t ∆(A) n (t. In contrast. Dn are the MC@NLO evolution kernels multiplied by Born matrix elements. D = ij. An important feature of Eq.sensitive to Φn is guaranteed by the fact that Eq. a slight modification is necessary.k Dij.4) is that it can be iterated to incorporate higher-multiplicity leading-order matrix elements into the prediction. however. This is a consequence of the fact that the two Sudakov form factors differ by their treatment of colour and spin correlations and it was discussed in detail in [12]. (2.2 Next-to-leading order matching . and In is the sum of integrated subtraction (A) (S) terms.k 1 and tn+1 = t(Φn+1 ) is defined in terms of Eq. as was detailed in [12]. In order to obtain this property when dealing with next-to-leading order matrix elements.MC@NLO In the MC@NLO matching method the following additional quantities are needed: • Squared real-emission matrix elements. 4 (PS) .7) with tN +1 = t(Φn+1 ) defined as above. Here. where the first line is the O(αs ) parton-shower result [20] and independent of Qcut .k . (2. summed (averaged) over final state (initial state) spins and colours and including symmetry and flux factors. It can be large if Qcut is either far from the collinear limit or sub-leading colour single logarithms are important. 3. µ2 n Q ) Θ(Qn+1 − Qcut ) O (Φn+1 ) . t ) = exp (A) − t dΦ1 Dn (Φn . Vn is the Born-contracted one-loop amplitude. 22].8) Note that ∆n implicitly depends on Φn . [12]. for n-particle processes. defined as (S) ¯ (A) B n (Φn ) = Bn (Φn ) + Vn (Φn ) + In (Φn ) (2. By incorporating full colour information in D(A) . Note that Rn (Φn+1 ) = Bn+1 (Φn+1 ). (2. Further. Both functions can be decomposed in terms of dipole contributions.5) dΦn+1 Bn+1 (Φn+1 ) − Bn (Φn ) Kn (Φn+1 ) ∆(PS) (tn+1 .

the higher-multiplicity tree-level result is “scaled up” by the local K -factor from MC@NLO in the soft region. The expectation value of an arbitrary. correspondingly are dubbed H-events [9.2.10) is determined by backward clustering. the hard remainder. without upsetting the logarithmic accuracy of the parton shower. Hence. as described in [5]. was independently proposed in [13] and [14]. 2. The improvement over results obtained by MEPS methods is the next-to leading order accuracy of the inclusive cross section and of observables sensitive to Φn . for configurations with a hard additional parton.9) dΦn+1 H(A) n (Φn+1 ) O (Φn+1 ) . The original algorithm. Note that the square bracket in the first line integrates to one.and H-events [9. µ2 Q ) Θ(Qcut − Qn+1 ) O (Φn+1 ) Bn (Φn ) (A) (2. 2. the n-parton phase-space configuration in Eq. This. we obtain kn (Φn+1 ) → (A) (A) ¯ (A) → Rn . Φ1 ) (A) ∆n (tn+1 . reflecting the probabilistic nature of the Sudakov form factor.1. Events generated n according to the first line are referred to as standard.e. and leading-order accurate for observables sensitive to Φn+1 . In order to see 5 .2 can easily be merged with higher-multiplicity event samples at leading order accuracy using the techniques described in Sec. for configurations with a soft additional parton. based on the POWHEG method [10. infrared-finite observable to O(αs ) in the MENLOPS method for MC@NLO is given by O = ¯ (A) dΦn B n (Φn ) µ2 Q 2 × ∆(A) n (tc . Correspondingly. 2. i. or S-events. Φ1 ) (A) ∆n (tn+1 . In both cases. In the limit Hn (A) dΦ1 Dn (Φn . implies that the total cross section reproduces the exact NLO result.e. (2. µ2 n Q ) Θ(Qn+1 − Qcut ) O (Φn+1 ) . while events generated according to the second line.The expectation value of an arbitrary infrared safe observable O to O(αs ) is then given by [9] µ2 Q O = + ¯ (A) (Φn ) ∆(A) (tc .11) + + (PS) dΦn+1 H(A) (tn+1 . together with equations (2. we have B n (Φn )/Bn (Φn ).10) instead kn (Φn+1 ) → 1.3. 12]. i. leading-order accuracy is maintained throughout the n + 1-particle phase space. (A) (A) 12].MENLOPS NLO-matched predictions as described in Sec. an MC@NLO prediction is next-to-leading order accurate for observables sensitive to the Born phase-space variables Φn . The square bracket on the first line is generated by a weighted parton shower. In contrast to the MEPS method. Rn (Φn+1 ) (2. This extension will be the topic of Sec. µ2 n (Φn+1 ) ∆n Q ) Θ(Qcut − Qn+1 ) O (Φn+1 ) (A) dΦn+1 kn (Φn+1 ) Bn+1 (Φn+1 ) ∆(PS) (tn+1 . which will be discussed (A) ¯ (A) in Sec. and it is left untouched in the hard region. µQ ) O (Φn ) + tc It is motivated by the behaviour of the underlying MC@NLO event sample in terms of S. while the terms dΦn B and dΦn+1 Hn correspond to fixed-order events. which requires the introduction of the local K -factor (A) kn (Φn+1 ) = ¯ (A) B n (Φn ) Bn (Φn ) 1− Hn (Φn+1 ) Rn (Φn+1 ) + Hn (Φn+1 ) . µ2 Q ) O (Φn+1 ) Bn (Φn ) (A) (2. µ2 ) O(Φn ) + dΦn B n n Q tc dΦ1 Dn (Φn . 23]. This prediction is next-to-leading order accurate for observables sensitive to Φn and leading-order accurate for observables sensitive to Φn+1 . 2. but it cannot be extended to higher parton or jet multiplicity.6) and (2. The method aims to maintain the full NLO-accuracy in the n-jet phase space and the LO-accuracy in the (n + 1)-jet phase space. In this publication we extend the method to MC@NLO. however. The key advantage compared to a pure NLO-matched prediction is that final states of higher jet multiplicity are treated as in the MEPS approach. In the limit Hn → 0.3 Combining NLO matching and LO merging . 4.7).

the MENLOPS result will be as well. In other words. O(αs L). Parton showers which are 3 extended to full NLL accuracy may become available in the future. the goal is to describe every jet observable at next-to leading order in the strong coupling constant. The method is constructed such that it is next-to-leading order accurate for observables that are sensitive to both Φn and Φn+1 Θ(Q − Qcut ). equation (2.9).12) with O MC@NLO given by (2. it is given by O corr = dΦn+1 Θ(Qn+1 − Qcut ) O(Φn+1 ) ∆(PS) (tn+1 . µ2 n Q) × (A) ¯ (A) ∆n (tn+1 . Omitting the obvious phase space arguments of the different matrix element contributions. and thus showing the desired property.that this indeed is the case. (A) 3 Merging at next-to leading order The previous section sets the scene to introduce a new prescription. It thus remains to show that the correction term does not introduce unwanted terms of higher logarithmic order. µ2 B n Q) (A) D 1 − (PS) Bn n ∆n (tn+1 . µ2 Q) (2. µ2 Hn B n Q) (A) Bn+1 − H(A) − D n n (PS) Bn+1 Bn ∆n (tn+1 . The logarithmic accuracy of the MENLOPS method therefore depends entirely on the logarithmic accuracy of the parton shower. µ2 n Q) × (A) ¯ (A) Hn B n 1− Bn Bn+1 + (A) (A) ¯ (A) ∆n (tn+1 .11) can be rephrased as follows: O = O MC@NLO + O corr . (2. which consistently merges multiple MC@NLO-matched event samples of increasing jet multiplicity. µ2 ) Q = dΦn+1 Θ(Qn+1 − Qcut ) O(Φn+1 ) ∆(PS) (tn+1 . and non-leading in 1/Nc .13) Since Dn is of O(αs L2 ) and because the ratio of Sudakov form factor is at most of non-leading logarithmic 2 3 2 order. including Sudakov suppression factors. Hence. statement is based on the logarithmic accuracy of currently available parton showers. while maintaining the logarithmic accuracy of MC@NLO for observables sensitive to Φn+1 . the overall contribution of this term is at most of O(αs L ). in which case the mismatch of O(α2 s L ) would be absent. If the parton shower is correct to NLL. Higher jet multiplicities exhibit the same accuracy as in the MEPS approach. 2 This 6 . the MENLOPS technique will not impair the accuracy of the parton shower itself.

7 . . µ2 B n Q) D(A) n (PS) Bn ∆n (tn+1 . i.3. tn+1 ) On+2 Bn+1 n+1 (A) (3. µ2 n Q) ¯ (A) B n (A) ∆n (tn+1 . (3. (PS) where again the obvious phase space arguments in the matrix element contributions and splitting kernels have been suppressed for better readability. the factor Bn /B n It is easy to show that next-to-leading order accuracy is maintained for observables sensitive to Φn+1 at Q > Qcut . tn+1 )∆(PS) (tn+1 . Eq.1) tn+1 × + (A) (A) ∆n+1 (tc . tn+1 ) (PS) (A) n+1 t − H(A) − n (A) ¯ (A) ∆n (tn+1 . µ2 n Q ) Θ(Qn+1 − Qcut ) On+2 + . 4. A convenient Monte-Carlo algorithm to generate ¯ (A) will be discussed in Sec. while all Sudakov factors ∆(PS) are generated by standard shower algorithms. 3 Additional contributions are at most of O (α2 L2 ) and thus do not impair the logarithmic or fixed order accuracy we intend s to prove.9). µ2 Q) dΦn+1 Θ(Qn+1 − Qcut ) On+1 ∆(PS) (tn+1 . µ2 ) On + dΦn B n n Q tc dΦ1 Dn 2 ∆(A) n (tn+1 . and where they have been moved to subscripts in the observables. Expanding (PS) the Sudakov form factor ∆n (t. tn+1 ) (PS) (A) . µ2 Q) (3. tn+1 ) ∆(PS) (tn+1 . tn+1 ) ∆n+1 (tc . 3. µQ ) Θ(Qcut − Qn+1 ) On+1 Bn (A) (PS) dΦn+1 H(A) (tn+1 . Taking into account only n + 1 parton final states the correction term this time is given by 3 O corr = ¯ (A) B n+1 = dΦn+1 Θ(Qn+1 − Qcut ) On+1 ∆n+1 (tc . which are dealt with in an iterative procedure as detailed in Sec.2. tn+1 ) ∆n+1 (tc .3) × D(A) 1− n Bn ∆(PS) (tn+1 .1 Definition of the MEPS@NLO technique We propose a method based on the following expression for the expectation value of an arbitrary infraredfinite observable O µ2 Q O = + ¯ (A) ∆(A) (tc .1) is rewritten as follows O = O MC@NLO + O corr .2) with O MC@NLO given by (2.2. In order to show the logarithmic accuracy of the procedure. 4. tn+1 ) On+1 + tc dΦ1 dΦn+2 Hn+1 ∆n+1 (tn+2 . µ2 n ∆n Q ) Θ(Qcut − Qn+1 ) On+1 µ2 Q + ¯ (A) dΦn+1 B n+1 Bn+1 1 + (A) ¯ B dΦ1 Kn n+1 n+1 t ∆(PS) (tn+1 . µ2 Q ) in the third line to first order and combining it with the square bracket 2 on the same line yields correction terms which are at most of O(αs ). The terms dΦn B n (A) and dΦn+1 Hn correspond to the fixed-order seed events. µ2 n Q) µ2 Q (PS) × Bn+1 1 + (A) ¯ B dΦ1 Kn n+1 ∆n+1 (tc . µ2 n Q) ¯ (A) B n+1 + Bn+1 µ2 Q − Bn+1 1 − dΦ1 Kn tn+1 ∆n+1 (tc . as discussed in ¯ (A) Sec. (3.e. . µ2 n Q ) Θ(Qn+1 − Qcut ) Dn+1 (A) ∆ (tn+2 . . where Q is the transverse momentum scale of the first emission. of parton n + 1. The square bracket on the first line and third line is generated by weighted parton showers. The dots indicate contributions from higher parton-level multiplicities.

7) dΦn+k+1 Θ(Qn+k − Qcut ) On+k+1 Hn+k ∆i i=n (PS) (ti+1 . for the case of the first additional emission. As before. ti ) 1+ Bn+k ¯ (A) B ti dΦ1 Ki (A) n+k ti+1 tn+k × + (A) ∆n+k (tc . it can be written as the sum of an MC@NLO-like expression acting on the (n + k )-parton Born configuration and a correction term. Thus the correction term n does not impair the formal logarithmic accuracy of the parton shower. ti ) . in general given by Qn+k = Q(Φn+k ).4). Those will eventually yield a contribution of O(αs L2 ). (A) 3. in general. ti ) . we now turn to the question how this can also be shown for the k th additional jet. In the formalism outlined here this is reflected by the Θ-functions involving the jet cut Qcut and the scale Q of the softest emission of a given Born-like (n + k )-jet configuration. tn+k ) On+k + tc D (A) dΦ1 n+k ∆n+k (tn+k+1 . tn+k )Θ(Qcut − Qn+k+1 ) On+k+1 Bn+k n+k (A) (3. 8 .and the logarithmic accuracy of the shower are maintained.Both terms in the curly brackets consist of one factor describing the emission of an extra particle. how NLO. if present. The first thing to be understood is that.2 Iteration for multijet events Having shown. follows a different construction paradigm. tn+k ) On+k+1 Bn+k n+k (A) (3.5) dΦn+k+1 Θ(Qn+k − Qcut ) Θ(Qcut − Qn+k+1 ) On+k+1 Hn+k ∆i i=n (PS) (ti+1 . while aiming at the same formal accuracy. In order to see the formal accuracy of this expression. the observable O will have support in different sectors by different jet multiplicities. The combination of virtual and real contributions in ¯ (A) B does not induce any logarithms spoiling the accuracy of the parton shower. by dropping the second Θ-function in (3. ∞ O excl n+k = j =n+k Oj Θ(Qn+k − Qcut )Θ(Qcut − Qn+k+1 ) . Dn and Bn+1 .6) where O MC@NLO n+k = n+k−1 ¯ (A) dΦn+k Θ(Qn+k − Qcut ) B n+k (PS) ∆ i (ti+1 . are due to sub-leading colour configurations stemming from the difference between ∆(A) and ∆(PS) . O excl n+k = ¯ (A) dΦn+k Θ(Qn+k − Qcut ) B n+k n+k−1 × ∆i i=n (PS) (ti+1 . tn+k ) On+k + tc D (A) dΦ1 n+k ∆n+k (tn+k+1 . and matching the showers to them. However. The factors multiplying these emission terms are at most of O(αs L). its authors start from the parton shower and correct its emissions with higher order matrix elements. (3. For such a configuration.1). It is worth noting here that the algorithm detailed in [17]. the respective expression for the (n + k )-exclusive jet part of the observable. ti ) × i=n Bn+k 1 + (A) ¯ B tn+k ti dΦ1 Ki (A) n+k ti+1 × + (A) ∆n+k (tc . like the approach presented here. let us define an (n + k )-jet inclusive expression of the observable. (3. O incl n+k = O MC@NLO n+k + O corr n+k . Rather than starting from the matrix elements. (3. these logarithms.4) is given by the suitably modified second part of Eq.

tn+k+1 ) ∆n+k+1 (tc .9) a procedure that has been used in LO merging for some time. (3.10) 4 In the case of the real-emission correction and the corresponding dipole subtraction terms we consider the underlying Born configuration instead. including the effect of truncated showering. the renormalisation scale should be determined by analogy of the leading-order matrix element with the respective parton shower branching history [5]. consequently. completely determined by the parton shower and thus correct by definition. it is sufficient to analyse the first emission off the (n + k )-jet configuration over the full phase space. tn+k+1 ) (PS) and the same reasoning already applied to Eq. µQ . (3. (3. in the Sudakov form factors. but only the accuracy of the hard matrix elements. The same scale should be used in Born matrix elements and real-emission matrix elements if they have similar kinematics.3) yields the desired result. tn+k ) .9). The same scale definition is used in the parton shower and. This scale has not been made explicit so far.The only difference with respect to the usual form of the MC@NLO expression in (2. see [19].9) is the term in the second line which encodes a veto on emissions into the jet region from intermediate lines with its O(αs )-part subtracted. 3. (3. ti ) ∆n+k (tn+k+1 . This can be n ) as [24] achieved if we define the renormalisation scale for a process of O(αs n n αs (µ2 R) = i=1 αs (µ2 i) . For a more detailed discussion. therefore the behaviour above this scale is defined by the parton-level result with next higher multiplicity. At the relevant order in αs . while the dependence on the resummation scale. Following the MEPS strategy. and in particular when the additional parton of the real-emission correction becomes soft or collinear.3 Renormalisation scale uncertainties The key aim of the MEPS@NLO approach presented here is to reduce the dependence of the merged prediction on the renormalisation scale µR . Here. however. The renormalisation scale uncertainty in the MEPS@NLO approach is then determined by varying µR → µ ˜R . By however extending the first emission above this cut and analysing the impact on On+k+1 we show that the two regions match as smoothly as the logarithmic accuracy of the parton shower dictates. Also.8) ti+1 − Bn+k+1 1 − ¯ (A) B n+k+1 + Bn+k+1 n+k dΦ1 Ki i=n t i+1 ∆n+k+1 (tc . one needs a definition which is independent of the parton multiplicity. 9 . the nodal scales µi found in the backward clustering on the Born-like configuration of a single event then enter the truncated showering. while simultaneously correcting for the one-loop effects induced by a redefinition in Eq. The second emission is. That is. To see this. clearly. Of course. This is a direct consequence of the fact that the parton-shower evolution is not improved in our prescription. In next-to-leading order calculations. µ2 i are the respective scales defined by analogy of the Born configuration with a parton-shower branching history4 . the (n + k + 1)-jet configuration. this correction term reads n+k+1 O corr n+k = dΦn+k+1 Θ(Qn+k+1 − Qcut ) On+k+1 − tn+k+1 ) 1− ∆i i=n ti (PS) (ti+1 . (PS) (A) × (A) Dn+k Θ(tn+k n+k−1 ¯ (A) B n+k + Bn+k i=n ti dΦ1 Ki (A) (3. the Born matrix element is multiplied by n αs (˜ µ2 R) 1− µ2 αs (˜ µ2 i R) β0 log 2 2π µ ˜ R i=1 n . The finding above shows that no terms appear due to the merging prescription that violate the logarithmic accuracy of the parton shower at and around Qcut . Note that only the dependence on the renormalisation scale is reduced compared to the MEPS method. tn+k ) ∆n+k (tn+k+1 . the phase space for this first emission is confined to the region below Qcut . which is employed in the computation of the hard matrix elements. The resummation scale dependence was analysed in great detail in [12]. remains the same. of course.

Evolution is restarted at t . The differential probability for generating a branching at scale t. the value is accepted with probability f (t)/g (t) [26].1) dΦ1 Kn (Φ1 ) Θ(Q − Qcut ) . • If the first emission is generated at scale t with Q > Qcut . 4 Monte-Carlo implementation In this section we describe the Monte Carlo implementation of the merging formula Eq.1) in SHERPA. this counterterm has the same functional form as the exponent (PS) of the Sudakov form factor ∆n (t.1 Generation of the parton-shower counterterm In addition. In our implementation of MC@NLO. 4. µ2 n Q) 1 1+ κ µ2 Q t (4. when starting from an upper evolution scale t is then given by t P (t. We can therefore use the following algorithm: • Start from an n-parton configuration underlying the n + 1-parton event at scale µ2 Q.2) The key point of the veto algorithm is. which is one of the basic ingredients to our method. (3. related to subleading colour configurations. integrated over z and φ. The related weights are applied analytically rather than using a hit-or-miss method.to generate the one-loop counter-term. by 1/κ.1). It is well known how to generate emissions according to Sudakov form factors with strictly negative exponent. µ2 Q ). Note that. the event weight is multiplied ¯ (A) (Φn+1 )/Bn+1 (Φn+1 ). • If no emission is produced. (4. much like the Sudakov form factor is computed in any parton-shower algorithm itself. one can still generate events according to P using an overestimate g (t) ≥ f (t). Secondly.2 Generation of the MC@NLO Sudakov form factor In this subsection we briefly recall an algorithm to compute MC@NLO Sudakov form factors [12]. a value t is generated as t = G−1 [ G(t ) + log # ]. The techniques needed to combine leading-order matrix elements with parton showers are given elsewhere [5]. that even if the primitive of f (t) is unknown. µ2 n Q) = 1 + κ µ2 Q t (PS) dΦ1 Kn (Φ1 ) Θ(Q − Qcut ) ∆(PS) (t . by construction. Assume that f (t) is the sole splitting kernel in our parton shower. They can thus be used to 10 . which cannot be interpreted in terms of no-branching probabilities and which are dealt with using a modified Sudakov veto algorithm [12. which is not necessarily an overestimate of f (t). t ) n Q ) ∆n ∆(PS) (t. while higher-order contributions remain the same. we have to deal with potentially positive exponents. the original n + 1-parton configuration is retained. if g (t) has a known integral.25]. (3. One can now introduce an additional estimate h(t). This simple algorithm allows to identify the O(αs ) counterterm with an omitted emission and to generate the correction term on-the-flight. Events will then be distributed as ∆(PS) (t. Firstly. however. µ2 (t. and implement a truncated parton shower with lower cutoff scale t. 4. t ) = f (t) exp − ¯f (t ¯) dt t . This leads to form factors larger than one. we now have to implement a method to generate the parton-shower counterterm on the third line of Eq. where κ = B n+1 • All subsequent emissions are treated as in a standard truncated vetoed parton shower.

Automated dipole subtraction [29] and the Binoth–Les Houches interface [30] are employed to obtain parton-level events at next-to-leading order with virtual corrections provided by the BLACKHAT library [31–34]. cf. . In order to match the customary notation we quote Ycut = (Qcut /Ecms )2 . tm ) = g (t) h(t) g (ti ) h(ti ) − f (ti ) .4% for MEPS@NLO. This guarantees that both acceptance and rejection terms are generated in sufficient abundance to reduce statistical fluctuations. Consequently.4% for MENLOPS. For each of the inclusive samples discussed in the following we generated 40 · 106 weighted events.and 4-jet NLO parton-level calculations combined with 5.3) can lead to negative weights. t1 . One can therefore guarantee a process-independent exponentiation of next-to-leading colour real-emission corrections in the MC@NLO. The sub-contributions in different jet multiplicities were automatically chosen according to their cross sections. . 5. Note that the hadronisation model in SHERPA [36] has been tuned in conjunction with the parton shower and leading order matrix elements.3) where the ti run over intermediately rejected steps. no leading colour approximation is made in the first step of the parton shower.2.6 CPU days (MC@NLO). 1. and 10. the predictions above the cut are remarkably stable and match the experimental data very well. The results presented here are at the hadron level.0 CPU days (MEPS@NLO) on Intel Xeon E5440 CPUs at 2. (4.7 CPU days (MENLOPS) and 2. which reflect the fact that sub-leading colour configurations are taken into account and that the a-priori density h(t) might underestimate f (t). 16].3% for MC@NLO. Leading-order matrix elements are generated with AMEGIC++ [27] and COMIX [28]. The cross section fraction of negative events was 1. In contrast to all other MC@NLO implementations. φ)-integral of these kernels. In order to implement an evolution using the MC@NLO kernels Dn /Bn we need to identify the function f above with the (z. 4. In the special case of e+ e− collisions discussed here this simplifies to the centre-of-mass energy. Within each jet multiplicity. The resummation scale is determined on an eventby-event basis by backward clustering as described in [5]. h (ti ) g (ti ) − f (ti ) i=1 m (A) (4. Note that Eq. The general-purpose event generator SHERPA sets the framework for this study [15. We are then free to choose the auxiliary function g on a point-by-point basis.and 6-jet at leading order. but a convenient way is to define g = C f .83GHz. . one should always choose the merging cut such that the analysis region is fully contained in the region covered by the NLO calculation of interest. The variation of results with Ycut in the region below and around Ycut is of the order of 10%. where C is a constant larger than one.3. 0. The above method guarantees that all subleading colour single logarithmic corrections to Bn are exponentiated. (A) 5 Results In this section results obtained with the MEPS@NLO method are presented for the case of e+ e− -annihilation into hadrons.absorb the negative sign of the MC@NLO kernels Dn /Bn .1 Choice of the merging scale Figure 1 shows the dependence of MEPS@NLO predictions for the Durham jet resolution on the merging scale Qcut . the number of H-events was statistically enhanced by a factor of 10 with respect to the S-events. It is therefore not surprising when deviations are found in observables that are sensitive to soft particle dynamics. Sec. This leads to a correction factor for one accepted branching with m intermediate rejections of w(t. rather than on the LO MEPS prescription that has been used so far in SHERPA. the related MC@NLO generator has been presented in [12]. The generation of 40 · 106 events needed 1. . A convenient choice of the function h will be the (z. 11 . φ)integral of the parton-shower evolution kernels Kn . The parton shower in SHERPA is based on Catani-Seymour dipole factorisation [35]. All results were generated using 2-. In the future this will necessitate a new tune of the hadronisation based on the NLO-merging outlined here.

= 91.6 2 3 4 5 6 7 8 9 10 11 − ln(y34 ) 1/σ dσ /d ln(y45 ) 10−1 10−2 10−3 10−4 10−5 1 .0 Ycut = 10−2.0 Ycut = 10−2.m.25 Sherpa+BlackHat MC/data 1 0 .6 4 5 6 7 8 9 10 11 12 13 − ln(y56 ) Figure 1: Experimental data from ALEPH [37] for the differential (n +1) → n jet rates with n = {2. 3. 12 .8 0 .75 Ycut = 10−2.6 1 2 3 4 5 6 7 8 9 10 − ln(y23 ) MC/data 1 . 2. left to right) at the Z pole (Ec.8 0 . 2.4 ALEPH data Ycut = 10−1.8 0 .2 GeV) 10−1 10−2 10−3 10−4 10−5 1 .75 Ycut = 10−2.8 0 .0 Ycut = 10−2.25 Sherpa+BlackHat 1/σ dσ /d ln(y34 ) Durham jet resolution 3 → 2 (ECMS = 91.2 GeV) 10−1 10−2 10−3 10−4 10−5 1 . the merging cuts have been indicated with dotted lines in the same colour in the ratio plot.0.0 Ycut = 10−2.2 GeV) Durham jet resolution 6 → 5 (ECMS = 91.75 Ycut = 10−2.25} .2 1 0 .4 ALEPH data Ycut = 10−1. To guide the eye.4 ALEPH data Ycut = 10−1.25 Sherpa+BlackHat MC/data 1 0 .2 1 .2 1 0 .6 4 5 6 7 8 9 10 11 12 − ln(y45 ) MC/data 1 .2 GeV) are compared with MEPS@NLO simulations with different values of the merging cut.75. Ycut = 10−{1.2 GeV) Durham jet resolution 4 → 3 (ECMS = 91. 5} (upper and lower panel.2 1 . 4.25 Sherpa+BlackHat 1/σ dσ /d ln(y56 ) Durham jet resolution 5 → 4 (ECMS = 91.1/σ dσ /d ln(y23 ) 10−1 10−2 10−3 10−4 10−5 10−6 1 .4 ALEPH data Ycut = 10−1.75 Ycut = 10−2.

5. which show a very good agreement between our simulation and data throughout. FS’s work was supported by the German Research Foundation (DFG) via grant DI 784/2-1. where we find a similar behaviour. Obviously in regions that are dominated by the parton shower. There the bands obtained from a scale variation in the matrix element regime are suspiciously small. which is equivalent to a variation of the corresponding resummation scale in analytical calculations.75 in the C -parameter. while the logarithmic accuracy of the parton shower is respected. that we focused on here. such a variation would give a more sensible estimate of theory uncertainties than a variation of the scales in the matrix element. the factorisation scale. Similar effects are observed in most event shape distributions in the hard region. The reduction of the scale uncertainty in the moments of the event shape distributions in particular is more than impressive. which we did not pursue here. there are. µQ . grant NSF–PHY–0705682. one could also vary the parton shower starting scale. In the MEPS@NLO sample we generate 2-. We have also implemented our formalism for the case of collisions with hadronic initial states [19]. where the 2 → 3 and 3 → 4-jet rates show significantly reduced uncertainties for larger y . 3.2 Comparison of approaches and their perturbative uncertainties In this section we compare the renormalisation scale uncertainties between the MENLOPS and the MEPS@NLO method. Another important source of uncertainty is the model for parton to hadron fragmentation. On the other hand. Figures 2 to 8 show the respective scale variations as bands around the central prediction with C = 1. This research is also supported in part by the Swiss National Science Foundation (SNF) under contracts 200020-138206. In addition. however. A significant reduction of the scale uncertainty is found for those observables. It is thus susceptible to variations of the renormalisation and. The results displayed here are exemplary for a far wider range of observables. consider the low-p⊥ regime of the differential jet rates yij . By explicit calculation. we have shown that our description maintains the higher order accuracy of the underlying matrix elements in their respective phase space range.3-. in addition to the variations outlined above. 2} and set Ycut = 2. Clearly. changes in the value of αs . for example in Fig. 2. (The LHC Theory Initiative). We therefore postpone this discussion to future work. This can be seen in particular in Fig. for T → 0. Acknowledgements SH’s work was supported by the US Department of Energy under contract DE–AC02–76SF00515. and in part by the US National Science Foundation. which is notoriously difficult to describe in fixed-order calculations. now shows a remarkably smooth behaviour due to the successful interplay of the different multiplicity contributions. 1.5. since in the small-y region both exhibit a comparable formal accuracy. if present. We choose µ ˜R = CµR with C ∈ {0. 6 Conclusions In this paper we have introduced a method for a consistent multijet merging at NLO accuracy for the case of e+ e− -annihilations to hadrons. − log yij → ∞. The most remarkable feature of our formalism is the greatly reduced uncertainty due to variations of the renormalisation scale. We have also analysed the impact of renormalisation scale variations in our new formalism. A final comment. is beyond the focus of this paper. concerning the evaluation of theory uncertainties by scale variations is in order here. MS’s work was supported by the Research Executive Agency (REA) of the European Union under the Grant Agreement number PITN-GA-2010-264564 (LHCPhenoNet). there are two sources of perturbative uncertainties: the one analysed here. A careful analysis of such effects. There.5. 13 . and by the Research Executive Agency (REA) of the European Union under the Grant Agreement number PITN-GA-2010-264564 (LHCPhenoNet). which discusses improvements of our ability to generate inclusive samples of events by increasing the formal accuracy of the matrix element part of the simulation. It is also worth pointing out that the typical Sudakov shoulder at C = 0. also uncertainties in the treatment of secondary emissions through the parton shower. As an example for this. which stems from the matrix element.and 6-parton final states at LO. obviously is true for the MEPS@NLO and the MENLOPS method. The MENLOPS sample only has the 2-parton final state at NLO and the remaining multiplicities up to 6 partons from tree-level matrix elements. or in parton distribution functions would have to be considered for a more complete assessment of such uncertainties. and it is clear that a variation of the resummation scale would yield larger uncertainties. and 4-parton final states at NLO and 5. which are sensitive to the NLO parton-level results. while the 4 → 5 and 5 → 6-jet rates do not. The same. of course.

.8 0 . .6 Sherpa+BlackHat 1 2 3 4 5 n Figure 3: Perturbative uncertainties in MENLOPS and MEPS@NLO predictions of thrust. .4 ALEPH data MePs@Nlo MePs@Nlo µ/2 .6 4 5 6 7 8 9 10 11 12 − ln(y45 ) MC/data 1 .2 GeV) 1/σ dσ /dT 10 1 1 10−1 10−2 10−3 1 .2 GeV) Durham jet resolution 4 → 3 (ECMS = 91. .2 1 0 .2 1 0 .2 ALEPH data MePs@Nlo MePs@Nlo µ/2 . 2µ MEnloPS MEnloPS µ/2 . . . 2µ Mc@Nlo Sherpa+BlackHat 1/σ dσ /d ln(y56 ) Durham jet resolution 5 → 4 (ECMS = 91.8 0 .4 ALEPH data MePs@Nlo MePs@Nlo µ/2 . . 2µ MEnloPS MEnloPS µ/2 . . . 2µ MEnloPS MEnloPS µ/2 .7 0.75 0 . Compared are the measurements for the event shape from ALEPH [37] and its moments from OPAL [38].85 0 .6 0. . 2µ MEnloPS MEnloPS µ/2 . . .2 GeV) 10−1 10−2 10−3 10−4 10−5 1 .6 4 5 6 7 8 9 10 11 12 13 − ln(y56 ) Figure 2: Perturbative uncertainties in MENLOPS and MEPS@NLO predictions of differential jet rates compared to data from ALEPH [37]. . Thrust (ECMS = 91. . 2µ Mc@Nlo 10−2 10−3 Sherpa+BlackHat 10−4 1 .4 1 . 2µ Mc@Nlo Sherpa+BlackHat MC/data 1 0 . .8 0.8 0 . 2µ Mc@Nlo 1/σ dσ /d (1 − T )n 10−1 Moments of 1 − T at 91 GeV OPAL data MePs@Nlo MePs@Nlo µ/2 .1/σ dσ /d ln(y23 ) 10−1 10−2 10−3 10−4 10−5 10−6 1 .8 0 .9 0.4 ALEPH data MePs@Nlo MePs@Nlo µ/2 . .2 GeV) 10−1 10−2 10−3 10−4 10−5 1 .8 0 . 2µ MEnloPS MEnloPS µ/2 .6 1 2 3 4 5 6 7 8 9 10 − ln(y23 ) MC/data 1 . . 2µ Mc@Nlo Sherpa+BlackHat MC/data 1 0 . . .2 1 0 . . .4 MC/data 1 0 .4 ALEPH data MePs@Nlo MePs@Nlo µ/2 .6 0 . .2 GeV) Durham jet resolution 6 → 5 (ECMS = 91.2 1 .95 T MC/data 1 .6 2 3 4 5 6 7 8 9 10 11 − ln(y34 ) 1/σ dσ /d ln(y45 ) 10−1 10−2 10−3 10−4 10−5 1 . .2 1 .8 0 . 2µ MEnloPS MEnloPS µ/2 .65 0 . 2µ Mc@Nlo Sherpa+BlackHat 1/σ dσ /d ln(y34 ) Durham jet resolution 3 → 2 (ECMS = 91. . 14 .

3 0. 2µ Mc@Nlo Sherpa+BlackHat 10−2 10−3 10−4 1 .8 0 .8 0 . .25 0 .8 0 .1 0. . 2µ Mc@Nlo 10 1 1 10−1 10−2 10−3 10−4 10−5 1 .35 0 .6 0 0. . . . . .6 Sherpa+BlackHat MC/data 1 . .2 1 0 . . .15 0 . .2 1 0 . 2µ MEnloPS MEnloPS µ/2 .4 MC/data 1 . . .8 1 C 1 2 3 4 5 n Figure 6: Perturbative uncertainties in MENLOPS and MEPS@NLO predictions of the C -parameter.6 Sherpa+BlackHat MC/data 1 . 2µ Mc@Nlo Sherpa+BlackHat 10−2 10−3 10−4 Sherpa+BlackHat 1 .8 0 .4 MC/data 1 .6 0 0 .05 0 .8 0 .2 1 0 .6 1 2 3 4 5 n Figure 5: Perturbative uncertainties in MENLOPS and MEPS@NLO predictions of wide jet/hemisphere broadening. . 2µ MEnloPS MEnloPS µ/2 .2 1 0 .2 1 0 . Wide jet broadening (ECMS = 91. .Total jet broadening (ECMS = 91.05 0 . . .4 0 . Compared are the measurements from ALEPH [37] and OPAL [38]. . .4 1 . 2µ MEnloPS MEnloPS µ/2 . Compared are the measurements from ALEPH [37] and OPAL [38]. .25 0 .2 GeV) 1/σ dσ /d BT n 1/σ dσ /d Bsum Moments of Bsum at 91 GeV 10−1 OPAL data MePs@Nlo MePs@Nlo µ/2 .2 GeV) 1/σ dσ /dC 1/σ dσ /d C n Moments of C at 91 GeV OPAL data MePs@Nlo MePs@Nlo µ/2 .2 0.4 BT 1 2 3 4 5 n Figure 4: Perturbative uncertainties in MENLOPS and MEPS@NLO predictions of total jet/hemisphere broadening. C-Parameter (ECMS = 91. 2µ Mc@Nlo Sherpa+BlackHat 10−1 10−2 1 . 2µ Mc@Nlo MC/data 1 10−1 10−2 10−3 1 . 15 .4 ALEPH data MePs@Nlo MePs@Nlo µ/2 .2 0. . 2µ Mc@Nlo 10 1 1 10−1 10−2 10−3 10−4 1 . . .2 GeV) 1/σ dσ /d BW n 1/σ dσ /d Bmax Moments of Bmax at 91 GeV 10−1 OPAL data MePs@Nlo MePs@Nlo µ/2 . 2µ MEnloPS MEnloPS µ/2 .4 ALEPH data MePs@Nlo MePs@Nlo µ/2 .8 0 .2 0 .15 0 . Compared are the measurements from ALEPH [37] and OPAL [38].6 0 0. . 2µ MEnloPS MEnloPS µ/2 . 2µ MEnloPS MEnloPS µ/2 .3 BW ALEPH data MePs@Nlo MePs@Nlo µ/2 .6 0 .4 MC/data 1 .1 0.2 1 0 .

5 0 MC/data MC/data 1 .8 0 .1 1 .0 0 .2 1 0 . .4 0 . 2µ MEnloPS MEnloPS µ/2 . 2µ Mc@Nlo 1 .5 1 cos(φKSW ) Modified Nachtmann-Reiter angle (parton level) ∗ )| 1/σ dσ /d| cos(θNR Angle between the two softest jets (parton level) 1 . .7 0 .5 0 . .0 .6 0 . 16 .5 0 0 .2 Sherpa+BlackHat 0 .8 1 ∗ )| | cos(θNR -1 . .5 2 1 . .6 1 . Compared are the measurements from ALEPH [37] and OPAL [38].6 0 . 2µ Mc@Nlo Sherpa+BlackHat 1/σ dσ /d Sn 10−1 Moments of S at 91 GeV OPAL data MePs@Nlo MePs@Nlo µ/2 . 2µ Mc@Nlo 10−2 10−3 1 .2 0 . .3 0 .4 0 . 2µ Mc@Nlo Sherpa+BlackHat -1 .9 S MC/data 1 .8 1 .1 1 .1 0 MC/data 1 . .9 0 0 .2 0 1 . 2µ MEnloPS MEnloPS µ/2 . .8 0 . .6 0 0 .5 3 2 . .8 0 .5 1 cos(α34 ) Figure 8: Four-jet angles using the Durham algorithm compared to data from OPAL [39].2 GeV) 1/σ dσ /dS 10 1 1 10−1 10−2 10−3 10−4 1 . .2 1 0 . . . .5 0 0 . .2 ALEPH data MePs@Nlo MePs@Nlo µ/2 .9 Korner-Schierholz-Willrodt angle (parton level) ¨ OPAL data MePs@Nlo MePs@Nlo µ/2 .2 0 .6 0 .1 0 .2 1 0 .7 0 .6 0 . 2µ MEnloPS MEnloPS µ/2 . 2µ Mc@Nlo OPAL data MePs@Nlo MePs@Nlo µ/2 .4 1 . 2µ Mc@Nlo Sherpa+BlackHat MC/data 1 .8 0 .4 0 . .1 1 . Bengtsson-Zerwas angle (parton level) 1/σ dσ /d| cos(χBZ )| 1/σ dσ /d cos(φKSW ) 3 . 2µ MEnloPS MEnloPS µ/2 . .5 0 .9 Sherpa+BlackHat OPAL data MePs@Nlo MePs@Nlo µ/2 .0 .0 0 .2 0 .4 0 .6 Sherpa+BlackHat 1 2 3 4 5 n Figure 7: Perturbative uncertainties in MENLOPS and MEPS@NLO predictions of sphericity.6 0 . .4 1 .8 0 .5 1 0 .6 0 .4 OPAL data MePs@Nlo MePs@Nlo µ/2 .9 0 0 .3 0 .0 0 . . . 2µ MEnloPS MEnloPS µ/2 .4 0 . . . . 2µ MEnloPS MEnloPS µ/2 . .4 MC/data 1 0 .0 0 .8 1 | cos(χBZ )| 1/σ dσ /d cos(α34 ) 0 .1 1 .Sphericity (ECMS = 91.

17 .TG and SH would like to acknowledge the Kavli Institute for Theoretical Physics (KITP) at UC Santa Barbara for its hospitality during the 2011 program ”Harmony of Scattering Amplitudes”. We gratefully thank the bwGRiD project for computational resources.

F. 18 . Schumann. Giele. JHEP 05 (2002). Hamilton. [12] S. S. 123. Hamilton and P. [20] S. F. Winter. H. Practical improvements and merging of POWHEG simulations for vector boson production. [18] S. 063. Sch¨ alicke. F. [hep-ph/0311263]. S. Higher-Order Corrections to Timelike Jets. NLO matrix elements and truncated showers. 024. Sch¨ onherr and F. Phys. Webber. Sch¨ onherr and F. F. [21] S. Kosower and P.1219 [hep-ph]]. 343–362. 046. Krauss.1127 [hep-ph]]. [2] L. Re. Sch¨ onherr. Maltoni. JHEP 1109 (2011). [hep-ph/0108069]. S. V. [8] L. 104. Gleisberg. Schumann and F. JHEP 02 (2004). JHEP 11 (2004). Kuhn and B. [hep-ph/0205283]. L¨ onnblad and S. H¨ oche. 007. H¨ oche. Mangano. [arXiv:0811. Moretti and R. F. [arXiv:1108. Matching NLO QCD computations and parton shower simulations. 019. F. S. Alioli. Improving NLO-parton shower matched simulations with higher order matrix elements.3072 [hep-ph]].4829 [hep-ph]]. Matrix elements and parton showers in hadronic interactions. Richardson and J. [arXiv:1109. Winter. JHEP 05 (2009). M. Pittau et al. M. L¨ onnblad. Hamilton and E. Gleisberg. [9] S. 049.1220 [hep-ph]]. [hep-ph/0204244]. Krauss.5030 [hep-ph]. Z. Lavesson and L. Krauss. Nason. Catani and M. Schumann and J. [4] F. A new method for combining NLO QCD with shower Monte Carlo algorithms. JHEP 12 (2008). Phys. JHEP 06 (2002). [arXiv:0905. [arXiv:0811. B632 (2002). Hirschi. R. B485 (1997). [arXiv:1009. JHEP 11 (2009). [6] K. F. F. M. F. Sch¨ onherr and F. A critical appraisal of NLO+PS matching methods. QCD matrix elements and truncated showers. [5] S.α. Siegert and J. [10] P. Catani.2912 [hep-ph]].1764 [hep-ph]]. JHEP 11 (2001). R. Sch¨ onherr and F. [7] W. 039. [hep-ph/0409146]. [11] R. [3] M. Multijet matrix elements and shower evolution in hadronic collisions: W b¯ b + n-jets as a case study.5399 [hep-ph]]. JHEP 1202 (2012).. JHEP 09 (2012). L¨ onnblad. [16] T. Frixione. Automating the POWHEG method in SHERPA. Nucl. [hep-ph/9605323]. Siegert. JHEP 06 (2010). H¨ oche. Krauss. [hep-ph/0109231]. S. H¨ oche. F. [hep-ph/0112284]. R. 054003. [arXiv:1111. Tully. L. S. Pittau. 038. 029. 291–419. Nucl. 048.References [1] S. R. Frederix. 040. Skands. A. Rev. JHEP 0208 (2002). [arXiv:1004.4622 [hep-ph]]. Phys. H¨ oche. [arXiv:1008. Webber. [13] K. a proof-ofconcept version. JHEP 03 (2012). Siegert. Matching Tree-Level Matrix Elements with Interleaved Showers. [15] T.0909 [hep-ph]]. Siegert. D84 (2011). Frixione and B. H¨ oche. Event generation with SHERPA 1. Krauss. M. Nason. [17] N. A. P. QCD matrix elements + parton showers. D. T. JHEP 04 (2011). Siegert. Krauss. Krauss. Siegert. M. 015.2126 [hep-ph]]. [arXiv:0903. arXiv:1207. M. A general algorithm for calculating jet cross sections in NLO QCD. aMC@NLO predictions for Wjj production at the Tevatron. [arXiv:1110. Krauss. Krauss. A modified CKKW matrix element merging approach to angular-ordered parton showers. K. [14] S. Correcting the colour-dipole cascade model with fixed order matrix elements. JHEP 08 (2011).1. [arXiv:1102. QCD matrix elements + parton showers: The NLO case. 070. [19] S. SHERPA 1. 056. H¨ oche. JHEP 02 (2009).5502 [hep-ph]]. Prestel. Extending CKKW-merging to one-loop matrix elements. Seymour. 053.

arXiv:1206. Abbiendi et al. [35] S. Commun. H. Phys. Bern. Forde. a new matrix element generator. J. C20 (2001). Measurement of event shape distributions and moments in e+ e− → hadrons at 91-GeV . Maˆ ıtre. Dittmaier and Z. F. Mrenna and P. Soff. Nason and G. Phys. Schumann and F. Eur. [26] T.. R. Nason and C. Berger. [34] C. C40 (2005). F. Febres-Cordero.. JHEP 05 (2006). Febres-Cordero. [hep-ex/0101044]. D. [hep-ph/0011222]. Automating dipole subtraction for QCD NLO calculations. Bern. [37] A. 19 . 038. S. [arXiv:0709. [28] T. [arXiv:1001. JHEP 02 (2002). J. [31] C. Forde.[22] S.3674 [hep-ph]]. D. D. J. 074002. [arXiv:0808. Phys. J. [arXiv:hep-ex/0503051 [hep-ex]]. Krauss. L.3501 [hep-ph]]. Phys. Dixon. [arXiv:1004.4180 [hep-ph]]. Gleisberg and F. Automated implementation of on-shell methods for one-loop amplitudes. Z. J.. 149–160. 181 (2010). Soff. [32] C. ALEPH collaboration.Phys. L. [38] G. Lett.209-GeV and a determination of alpha(s). H. MINLO: Multi-scale improved NLO. Kosower and D. JHEP 03 (2008). A. Winter. Comix. Maˆ ıtre. F. Lett. D78 (2008). B500 (2001). Dixon. Gleisberg.0: A Matrix Element Generator In C++. S. Ita. [24] K. J. 070. C35 (2004). Studies of QCD at e+ e− centre-of-mass energies between 91 and 209 GeV. 381–395. Forde. Siegert. Krauss. γ ∗ +3-Jet distributions at the Tevatron. 601–615. Next-to-leading order QCD predictions for W+3-Jet distributions at hadron colliders. Gleisberg and S. Bern. J. Trocsanyi. D80 (2009). A simultaneous measurement of the QCD colour factors and the strong coupling. JHEP 11 (2007). Precise Predictions for W + 4-Jet Production at the Large Hadron Collider. Phys. 501–523.3572 [hepph]. One-loop singular behaviour of QCD and SUSY QCD amplitudes with massive partons. Z. [27] F. Eur. [36] J. T. Phys. D. C53 (2008). [arXiv:0709. [arXiv:1009. L. Phys. A modified cluster-hadronisation model. 074036. [25] S. F. F. Rev. Krauss and G. D. Kuhn and G. [arXiv:0803. P.1659 [hep-ph]].4 physics and manual. Oleari. Ita. T. Dixon. Eur. Hamilton. Dixon. JHEP 12 (2008). Eur. [23] S. 092001. D. PYTHIA 6. H. Phys. C36 (2004). F. [arXiv:0907. Kosower and D. 044. T. Frixione. F. A. Phys. [39] G. OPAL collaboration.1027 [hep-ph]]. Berger. H¨ oche. 036003. Rev. Z. Ita. [arXiv:0912. D82 (2010). [hep-ph/0109036]. Phys. D81 (2010). [arXiv:0709. D. F. Febres-Cordero. Eur. 034026. Febres-Cordero. H. 106 (2011).1307 [hep-ph]]. 039. [33] C. Forde. [hep-ph/0603175]. Abbiendi et al. Phys. S. Maˆ ıtre.1984 [hep-ph]]. H¨ oche.2338 [hep-ph]]. 457–486. Heister et al. L. Zanderighi. Kosower and D.-C. Ita. Kosower and D. Rev. Berger. Hard photon production and matrix-element parton-shower merging. Sj¨ ostrand. Comput. D. J.J. P.. Gleisberg. Rev. Skands. [29] T. Matching NLO QCD computations with parton shower simulations: the POWHEG method. A parton shower algorithm based on Catani-Seymour dipole factorisation. A proposal for a standard interface between Monte Carlo tools and one-loop programs.2092 [hep-ph]]. Z. Bern. A. Catani. OPAL Collaboration collaboration. Berger.2881 [hep-ph]]. [hep-ph/0311085]. 026. AMEGIC++ 1. A. Gleisberg. Schumann and F. 1612–1622. Maˆ ıtre. Rev. [30] T. 287–316. Krauss. F. Next-to-leading order QCD predictions for Z. Binoth et al.