Quantifying the performance of jet definitions for kinematic reconstruction at the LHC

TL;DR

This paper evaluates various jet algorithms at the LHC using a new, purely observable-based method, identifying optimal jet definitions for different energy scales and jet types to improve kinematic reconstruction.

Contribution

It introduces a novel strategy to assess jet definition performance solely with physical observables, independent of distribution shapes and Monte Carlo references.

Findings

Small jet radius (R≈0.5) is optimal for moderate energy quark jets.

Larger radii (up to R≈1) benefit gluon jets and TeV-scale events.

Results guide the choice of default jet definitions for LHC analyses.

Abstract

We present a strategy to quantify the performance of jet definitions in kinematic reconstruction tasks. It is designed to make use exclusively of physical observables, in contrast to previous techniques which often used unphysical Monte Carlo partons as a reference. It is furthermore independent of the detailed shape of the kinematic distributions. We analyse the performance of 5 jet algorithms over a broad range of jet-radii, for sources of quark jets and gluon jets, spanning the energy scales of interest at the LHC, both with and without pileup. The results allow one to identify optimal jet definitions for the various scenarios. They confirm that the use of a small jet radius (R\simeq 0.5) for quark-induced jets at moderate energy scales, O(100 GeV), is a good choice. However, for gluon jets and in general for TeV scales, there are significant benefits to be had from using larger radii, up to R\gtrsim 1. This has implications for the span of jet-definitions that the LHC experiments should provide as defaults for searches and other physics analyses.