We now look at the effect of incrementally adding the suggested cuts, as well as the effect of using the additional cuts from Weihong's thesis that are proposed to be no longer used.

The Cuts

The cuts, in order of application, are (quoting from McPi0CutOptSelectorAddative.h)

 * 0. Merge the pi0s if delta R < 0.05 (only applied to rec, not gen)
 * 1. Epre1 < 40 MeV
 * 2. Esmd / Etow > 0.006
 * 3. pi0 eta in [1.086, 2]
 * 4. Vertex Z in +/- 150 cm
 * 5. at least 1 tower in event with E_T0 > 1.5 GeV
 * 6. photons' E_T0 > 1.5 GeV
 * 7. Number of hits in the 6-18 used towers < 5
 * 8. both photons in the same sector
 *
 * Note, the cuts are addative, so cut #1 also includes cut #0, and
 * cut #2 also includes cuts #0 and #1.  Only cuts #1, #2, #3, and #6
 * are applied to the generated photons.
 *
 * All the above cuts also include cuts on the pi0 pT range (inherited
 * from the parent) and the photon energy.  The photon energy cut is
 * 1.5 GeV (IU), 2.0 GeV (TSIU) and 2.5 GeV (TSP).
 * 

In the above E_T0 refers to the E_T using the detector eta (i.e. the origin not the vertex position).  Note: cuts #1 and #2 are applied to both photons in the pi0 candidate.  Whether this is taken care of correctly in the case of merged pi0s is under investigation.  The photon energy cuts and pT cuts are the same as from my last blog.

Results and Conclusions

The plots are organized just as in my last blog.  The x-axis now corresponds to which of the above cut levels is the last applied.  Based on the analysis, it seems wise to stop after cut #2.  However, for better comparison with theory, it might be better to stop after cut #3.  Naively, it is puzzling that the efficiency drops because of cut #3, but most likely it is due to there being more pi0s at lower eta and that the vertex is displaced in the negative z-direction.  Neutral pions are also more easily reconstructed at lower eta, as the seperation between the photons is larger.  This suggests looking into evaluating the eta-dependence of the efficiency, as well as looking at the eta range to consider whether it'd be better to use cut #3 but use a different range than 1.086 to 2.0.