Studies on the improvement of the matching uncertainty definition in top-quark processes simulated with Powheg+Pythia 8

Top-quark processes, including pair production, single-top production and production in association with electroweak bosons, are modelled by the ATLAS Collaboration by matching the hard-scatter matrix element calculations of next-to-leading order Monte-Carlo generators with a parton shower generator. This document presents a comprehensive study of the uncertainty related to the matching procedure and a new strategy to evaluate this uncertainty. The new approach is based on the Pythia 8 parton-shower matching parameter $p_\mathrm{T}^\mathrm{hard}$. It is designed to surpass the previous method, which involved comparing two generator setups to cover the uncertainty. The old method entangled all differences between the two setups in a single uncertainty while the new prescription implements a focused uncertainty that avoids double-counting with other uncertainties on the modelling of the top processes. Two other effects, previously covered by the matching uncertainty, are studied in parallel: the mismodeling of the top-quark transverse momentum and the top-quark decay lineshape. The first is addressed through a reweighting to the NNLO analytic calculation while the second is studied using two different calculations for the top-quark decay. While the first study leads to the definition of an uncertainty prescription, in the second case it is concluded that comparisons with more advanced generators are needed to define an uncertainty on the lineshape. The samples generated to define the new prescriptions are compared with the setups previously used using unfolded data at a centre-of-mass energy of $\sqrt{s}= 13$ TeV and Monte-Carlo simulations of both top-quark pair and single top-quark processes.