At the PWG meeting a few weeks ago, people expressed concerns about the high-mass tail from the pion peak that, for higher PT bins, is underestimated by the pion peak simulation.  This is shown (in the most egregious case) below (see especially, the left-hand plot between .2 and .5):

It is clear, at least at high-pt that the aggregate MC model does not fully reproduce this 'bleeding edge'.  Naturally the question arises of how sensitive the cross section measurement is to this tail.  One way to probe the sensitivity is to vary the invariant mass cut around the peak position.  For the preliminary result, I used a 3 sigma cut calculated by fitting the mass peak to a gaussian.  I can also calculate the cross section using a 2 sigma and 4 sigma cut.  These cuts are shown below...

This plot shows a closeup of the pion peak for each of the 9 bins, with vertical lines showing the three different mass windows.  For most bins, the 2 sigma cut completely excludes the tails and the 4 sigma cut includes a large portion of the tails.

I can then compare the different windows to the nominal window and the smaller (larger) windows.

The largest % difference is 13.5%, in the 8th bin.  For the most part the higher pt bins can be off by ~10% for a large mass window.