Monte Carlo Tuning for $e^+e^-\to$ Hadrons and Comparison with Unfolded LEP Data

Abstract: We perform two tunes of the SHERPA Monte Carlo generator for the generation of $e^+e^-\rightarrow\mbox{hadrons}$ using the publicly-available LEP analyses in Rivet. In each of these tunes, we generate events at $\sqrt{s}=91.25\mbox{ GeV}$ using matrix elements for final states containing up to six partons. In the first, "LO" tune, matrix elements for all final states are generated at leading order; in the second, "NLO" tune, matrix elements for final states with up to four partons are generated at next-to-leading order using BlackHat, while those for states with five and six partons are generated at leading order. The tunes are accomplished using Professor, and comparisons with unfolded LEP1 and LEP2 data are produced with Rivet. We also compare the data with events generated with KK2f interfaced to PYTHIA using the standard ALEPH tune. We find that both SHERPA samples show improvement relative to KK2f for observables related to four-jet final states, while all three samples produce comparable results for event-shape variables. Overall, the agreement with data is best for the LO tune. We provide our tuning parameters and many data-Monte Carlo comparisons.