The point of pass 1 is to make sure all issues are brought to the table, and that the code structure is ready to get the final result.  Pass 1 is now complete, and the next step is to improve the major problem areas and solve the open issues.

Choices made for Pass 1

• Reconstruction Algos: standard IU except consider three different SMD clustering algos, IU, Morhac and TSP.
• Background Subtraction: fit a skewed Gaussian (signal) plus a linear function (background) in the region 0.08 to 0.3 GeV in invarient mass.  Integrate the number of pi0s (after background subtraction) using the range 0.1 to 0.2.
• Efficiency: IU method, where one computes the efficiency only considering the generated pT bin, and ignoring whether the reconstructed pT is close to the generated pT.
• Formulas used for computing cross section and spin asymetries are posted in this blog.

Open Issues

• Double check computation of the statistical uncertainty
• Reducing various sources of background, specifically falsely split clusters & conversions
• Best way to subtract the background
  • Improve estimate of combinatorial background and determine better way to use it (see this talk for current status).
  • Consider Pythia's background
• Unfolding the yield per pT bin to correct for pT bin migration
  • Do it using the naive 1D migration.
  • Does it need to be invarient mass dependent, or include some other variable?
• Reconstruction algos
  • Evaluate agreement between reconstruction algos on an event by event level
  • Determine some answer regarding the (current) major differences in the estimated cross section and asymmetries
  • Ensemble methods, combining several different reconstruction codes
  • Non-clustering methods, to reach higher pT (Method of Moments and the conversion method)
• Estimate systematic uncertainties
• Review cuts

Description of the Results

Results for each of the algos are given in the attached PDFs.  A subset of 51 runs of the 2006 data was used.

• Pages 1-2: the lumi-weighted yield per pT bin
• Page 3: the efficiency matrix
• Page 4: results