E852 experiment

Analysis of Eta Pi0 system with the decay Eta -> Pi+ Pi- Pi0

 

********************************************

Draft 5_3

Date: 11 Jul 2006 From: Gary Adams for TRC

Subject: TRC comments and response

Vladimir and Ludmila: The trc discussed your analysis on Monday (draft 5.3s). We tried to identify what we thought could be published with little additional work. The recommendations below constitute our best estimate for reaching that goal. Of course you should feel free to present alternative plans if you disagree. We appreciate all of the work you have done so far.

1) We feel that it is not yet possible to determine what the conclusions of the paper will be, but in the interest of reaching a conclusion soon, some reduction in the scope of the paper is advised. Neal has agreed to help with the writing job when you and the trc agree on the outline.

2) Further work on the t' dependence of the pwa is not recommended. Suh-Urk's tests are not likely to affect the outcome for the large unnatural parity waves.

3) We do not feel that the MDPWA method should be included in the paper. We still have many questions about the method and answering those would require more analysis. We are also concerned that this method is unstable with respect to the pi1 width. Fixing the a2 parameters changes the pi1 width and we don't understand why.

4) We are confident that the PWA+MDF method can be published. Some more MDF analysis needs to be done, but not much. Fixing the parameters of the a2 in the MDF is the best way to finish the analysis and it is easy to motivate in a publication. You have already done much of the required work. In her last email Ludmila reported pi1 parameters of M=1273+-17 +47 -13 and W=373+-55 +10 -150 using RMS values for the systematic errors. These need a little more study, as outlined below.

5) The best values for the a2 parameters have not been used in the MDF. Using the PDG mass is OK but the width needs to be broadened to account for resolution. We don't have a good simulation to determine this but if we start with the eta pi- result (118 MeV) then we can make a pretty good estimate. We suggest fixing the a2 width at 120 MeV. To add this into the study of sys errors you should include 110 MeV and 130 MeV. Your sys study already uses 118 MeV so some of this has already been done.

Response: See Note 8

6) We still have some questions about the MDF analysis method. For example, on page 4 of Note2 you report chisq/dof around 1.2 using the average of the ambiguous solutions, but the fits to individual random selections of PWA results (fig 5 page 9) give much larger chisq, with none as low as 1.2. Why are these so different?

Response: See Note 8

7) As Ludmila has pointed out, the distribution of chisq values (page 9) has a long tail on it and this affects how we report the sys errors. We would like to see scatter plots of pi1 mass and pi1 width vs. chisq. Also, we would like to know what these look like, and what fig 5 looks like, if you reject fits that have large chisq (>2 sigma). These would correspond to fits that match the resonance hypothesis well. We note that the mass and width in Table 1 of the paper do not match the peak positions or averages in fig 5 page9 of note 2. This means that the resonance parameters are correlated with chisq so we need to document that. Lastly, we would also like to see scatter plots of pi1 mass vs. pi1 width, with and without the chisq cut.

Response: See Note 8

8) The scatter plots and histograms requested in item 7) suggest an alternative way to determine the pi1 resonance parameters. The average values, or the most probable values, could be taken from the frequency histograms after a chisq cut. The random errors could be taken from the widths of the distributions. This would avoid averaging the amps before the MDF. The answers should be similar to your previous method but this method would be easier to explain. We would like to see these results for comparison.

Response: See Note 8

9) It looks like the label on fig 2c in the draft is wrong. It should be D-P, not P-D.

Response

Doen

Response