Hi Ric, thanks for comments. > * the resolution on deltaM is quite poor. This might be just normal, but > can you quantify the resolution (it is about 0.8 MeV usually, when > everything is reconstructed). here we have two factors that worsen the resolution: 1) missing tracks 2) presence of extra bkg photons and missing photons Here you see a further zoom of my previous plots. http://www.slac.stanford.edu/~daniele/vub/deltamstudy2/b0cocktail.ps It is clear that there are two components. The first one with the ~right resolution, the second with a bad resolution (that probably contains the pathologies I previously mentioned). To confirm this I redid these plots with the additional request of no photons in the recoil. http://www.slac.stanford.edu/~daniele/vub/deltamstudy2noneu/b0cocktail.ps You see that the resolution is really improved now. (the fact that this cut keeps only the good events can be seen in the M(X - 1pi) plot. http://www.slac.stanford.edu/~daniele/vub/deltamstudy2/b0cocktail_MD0.ps http://www.slac.stanford.edu/~daniele/vub/deltamstudy2noneu/b0cocktail_MD0.ps ) > * do you understand why the signal has a peak (which is actually much > narrower than the background one)? I noticed these peaks also yesterday when I produced the first plots (yes, because it seems to be more than one peak). In http://www.slac.stanford.edu/~daniele/vub/deltamstudy2/sigocktail.ps you see again the usual plots for the signal. The peak is around 150 MeV and there is also a second peak around 170 MeV. It is much more clear once with divide the sample in B0 and ChB. http://www.slac.stanford.edu/~daniele/vub/deltamstudy2/sigocktail_B0.ps http://www.slac.stanford.edu/~daniele/vub/deltamstudy2/sigocktail_ChB.ps In the ChB the two peaks are evident. If I try to plot the same quantity for events with no neutrals in the recoil I get the following: http://www.slac.stanford.edu/~daniele/vub/deltamstudy2noneu/sigocktail.ps and then the particle obtained after subtracting a charged pion does contain neutrals. > * from your numbers it does not look like a good reason to cut on it ... > I mean, for B0 it could be interesting. I report again the efficiencies - B0 cocktail (b->clnu) Eps (no cut on MX) = 76% Eps (MX<1.6) ~90% - Signal (b->ulnu) Eps (no cut on MX) = 98.4% Eps (MX<1.6) = 98.5% I remind you that these efficiencies are calculated on the sample obtained applying all cuts. The only problem could be the estimate of the efficiency on the background and the effect of this cut on the background shape. Daniele