The shape of the invariant mass distribution using Pythia data, 2006 Geometry, QCD events.  Only pi0 candidates near a generated pi0 are considered, thus naively one could call all the data "signal."  The results are presented as a set of slides, with a separate set for each of the three algorithms currently under investigation.

Note, while the the initial mass distributions (slide 2) are quite different for the various algorithms, under certain cuts (slide 3, also slides 28-31, among others) the distributions look very similar.

The conclusions, as stated on the last slide, are as follows:

• Problems due to falsely split clusters still present
• Other newly (re)discovered issues
  • pi0 candidates from a true p0, but at least one photon converts in a manner that the general direction of the pi0 is still reconstructed but sin\alpha/2 resolution is poor.
  • Resolution on E_pi0 is poor
  • Exists almost no correlation between reconstructed and generated values of Z_gg
• Each of the above items can be further investigated, to either limit the effect or to correct for the effect.