Now that we are confident that the efficiency we're calculating from Pythia6 is reasonable, and now that we are confident that our parameterized response is implemented correctly as well, we want to check how the resulting response matrices compare.  Additionally, we're stilling trying to get a better understanding of the tracks in the Run9 dijet embedding sample (Pythia6) that are not matched to a generated particle.  Some of the previous investigations into these tracks can be found here:

https://drupal.star.bnl.gov/STAR/blog/dmawxc/update-06072019-run-9-embeddingmatched-vs-unmatched-track-comparison-including-those-mat

One question to ask would be how much do these unmatched tracks affect our response matrix.  If their contribution is negligible (with respect to Pythia8), then we don't really have to worry about them; since we're working in pp, we would expect the effect of pile-up to be quite small.  The plots below compare the response matrix extracted from Pythia6 via the embedding process (which includes the unmatched tracks) and the response matrix created in Pythia8 using the parameterized efficiency (which does not include unmatched tracks).

The following parameters were used in creating these matrices:

• Using real (particle-level) charged hadron (Pythia6) or pi0 (Pythia8) triggers
• eTtrg = (9, 20) GeV/c
• |etaTrg| < 0.9
• |vZ| < 55 cm (for Pythia6)
• Considering only charged recoil jets
• R = 0.2, anti-kT
• pTconstituent = (0.2, 1000) [particle-level jets] or (0.2, 30) [detector-level jets]
• |etaConstituent| < 1
• |etaJet| < 1 - R
• areaJet > 0.05

We should expect some differences at very, very low pT (< 1 GeV/c or so) due to differences in the underlying event between the simulation, and the presence of unmatched tracks should have a (hopefully very small) impact on the Jet Energy Resolution in the Pythia6 case.  Indeed, we see both of those things, but on the whole the response matrices agree well.