Attached are updated plots for the data/MC comparison.  A bug was fixed in reading out the results of the simulated trigger, so that now the L2 part of the trigger is also included (the bug caused only the L0 portion to be utilized).  Additionally, both the data and Monte Carlo now have a cut on the simulated trigger, with no hardware trigger cuts begin used.  The thresholds for the simulated trigger are set to be the thresholds from the hardware triggers.  The L2 thresholds are 3.7 GeV (TP) and 5.2 GeV (HT).  The L0 thresholds are those of set 'f' on the last table in this other blog.

Additionally, I've made an additional cut on a "mock-up" L2 trigger, where I look for a high tower with ET > some HT threshold and require the sum of the ET of the 3x3 patch of towers surrounding it (including the HT) to have ET > some TP threshold.  I'm using a lower threshold of HT = 2 GeV and TP = 4 GeV, intentionally set below all hardware L2 thresholds used in 2006.  For Monte Carlo, I'm using HT and TP thresholds of 3.7 GeV and 5.2 GeV, respectively, to match the L2 values.  The exact values for data and Monte Carlo to get a good comparison are still being optimized, but already the comparison is much improved over last week. Other various HT and TP thresholds were tried in the "mock-up" L2 trigger.  It is unknown why the thresholds here seem to be lower than I'd expect based on the L2 thresholds in the simulated trigger.

The vertex is still a bit off, but much improved.  The mass comparison is also much improved, with possibly no longer a descrepancy at low mass (other than the admixture of conversions, other background, and signal are off, as usual).  Other variables seem in fairly good agreement, as before.