As requested during the SPIN meeting yesterday, I've divided my max tower ET plots into reconstructed pi0 pT bins.  I've also included the max ET (tower) / ET (3x3) plot (with ET in 4-6 GeV) and the max ET (tower) / ET (3x3) plot vs max ET plots.  In making these plots, I've also raised the L0 threshold a second DSM value and raised the L2 thresholds to 3.9 GeV (HT) and 5.5 GeV (TP).  The L0 and L2 are now both 10% higher than the max value for any of the hardware trigger settings.  The following uses a subset of the data and all the Pythia. [It seems this is a misprint: the L2 threshols were 4.2 GeV (HT) and 6.0 GeV (TP) -- sgliske, Jan 11th, 2013]

To fix the normalization, I compared the distrubution of the pi0 pT, which is the following figure.

The normalization is fixed in the region 8-16 GeV.  The spectrum matches well except at the lowest pT.  Recall that the 5-6 GeV bin is used in the unfolding, but is not presented in the final cross section result.  Thus, overall the proposed range (6-16 GeV), the agreement is good.

The other plots are attached as pdf files.  The first slide is the full, 4-16 GeV spectrum, and the following slides are each reconstructed pi0 pT bin.

Discussion and Systematic Error Estimation

The comparison in the pT bins which are shown (6-16 GeV) is actually quite good.  The troubles in the lowest pT bin is likely due to false cluster splitting, and thus represent artifacts of higher pT events.  The max ET plots and ratios for pi0 pT > 6 GeV show good agreement, except possibly at the larged ET.  This region has already saturated the trigger, so the slight difference is negligable (since all other algorithms but the trigger sum over the 3x3 and do not consider just the high tower).  The ratio vs ET plots show nice trends with increasing pi0 pT, but not show little variation of differences in shower shape between data and Monte Carlo for pi0 pT > 5 and max ET < 5.

Note: the major change in these plots vs. those shown yesterday is the requirement of at least one reconstructed pi0 pair with pT > 4 GeV.  The problems seen yesterday predominately occur in data with no reconstructable pi0 with pT > 4 GeV.

While I was waiting for the runs to finish, I determined a correctio factor (which could either be used to correct or to estimate of systematic uncertainty) based on the ratio of the integrals of the max ET plots above or below max ET of 5 GeV.  However, the agreement is now quite good, and the correction factor is small for all relevant pT bins (pT>6).  The factors are

 4-5 2.00456
 5-6 1.16786
 6-7 1.07328
 7-8 1.04166
 8-9 1.04165
 9-10 1.01484
10-12 1.03516
12-16 1.02484

These corrections are far below the current uncertainties.

Other Data/MC Comparison Plots

I've also attached the latest data/MC comparison plots for other variables.  Again, the agreement is pretty good for pi0 pT > 5 or 6 GeV.  [2:20pm CDT, Oct 15--I realized I attached the wrong set of plots for the other kinematic variables.  I fixed it now.  The plots with the newest trigger threshold should have an "-h-h" in the name, not a "-g-g".]

Eta, g-g
Vertex Z, g-g
Z-gamma-gamma, g-g
M-gamma-gamma, g-g
E-gamma-gamma, g-g