At the centrality meeting on 11/16/2021, people asked for the amount of pileup removed by my proposed pileup cuts in each Vz bin. I was also asked to investigate increasing the upper cut in refMult vs. BTofMatched space because it appeared that the upper cut was eating into the distribution for large |Vz|. Below is my comparison. It seems that the actual number of pileup events rejected in the current method is nearly independent of Vz.