As part of the systematics for the analysis, we would like to vary the tracking efficiency by 5% or so.  Since I'm working with the Run 9 dijet embedding, the tracking efficiency is set by the Geant simulation of STAR.  First things first, some details about the Run 9 tracking efficiency can be found here:

https://drupal.star.bnl.gov/STAR/blog/dmawxc/update-04112018-run-9-pp-efficiency-different-binning-0

To adjust the efficiency up or down, I throw out additional particles at the particle or detector level respectively.  In other words, the efficiency is pulled down as so: if a track passes the usual track QA cuts, a random number is (uniformly) generated and 5% of the time the track is thrown out.  And the efficiency is pulled up by randomly throwing out 5% of the particle-level tracks (since dividing by 0.95 is the same as multiplying by 1.05).  Below are the results of this procedure.

This will be implemented in constructing our jet distributions by randomly removing 5% of particles from the jet-finding process at the particle or detector level.