 |
 |
Hi, I have been looking a bit more into the effect reported here by Daniele. There seems to be a significant jump, mostly a drop in the B0 background MC, between the selection gam=0/trk=3 and the selection gam=6/trk=6. The differences between the two selections are : gam 0 : energyGam[i] >= 0.08 trk 3 : goodtrackveryloose + acc gam 6: energyGam[i] >= 0.08&& energyGam[i]<4. && lMomGam[i]>0.05 && lMomGam[i]<0.5 && s9s25Gam[i]>0.9 && acc; trk 6 : goodtrackveryloose + require Dch hits if Pt>200 MeV (note: no acceptance cut) No unmatched neutral cluster removal in either selection. The basic difference is due to the cut on the lateral momentum (between 0.05 and 0.5) and the S9/S25. I have compared the MC shapes for gam=0,trk=3 (dots) and gam=6,trk=6 (histo) both in MC http://www.slac.stanford.edu/~rfaccini/phys/vub/comp_110602B0-03/-comp.html and in data http://www.slac.stanford.edu/~rfaccini/phys/vub/comp_110602B0-03data/-comp.html I would say that the quantities related to the tracks undergo minimal changes, while the neutral related ones change quite a bit. We therefore decided to factorize the problems and run only with trk 6 and gam=5 i.e. the same selection without the Lat cut, which might be particularly different between data and MC. This turned out to yield, for the B0 BRBR = 0.0393843 +- 0.0069899(stat) +- 0.00342245(MC stat) which is to be compared with BRBR = 0.0274513 +- 0.00629619(stat) +- 0.00355559(MC stat) in g00 and BRBR = 0.037327 +- 0.00518964(stat) +- 0.00305673(MC stat) in g03 It looks like the jump is caused by the other cuts ( energyGam[i]<4. or s9s25Gam[i]>0.9 && acc which look quite innocuous. the data MC agreement is in http://www.slac.stanford.edu/~rfaccini/phys/vub/comp_110602B0-trk7gam5/-comp.html comments by voice, since the meeting has started ... Ric P.S. in checking the options I realized that neither the trk=6 nor the trk=7 option have the tracking acceptance turned on. I then committed a version of the code with the acceptance on (I do not expect it to be a dramatic effect, but it was a mistake) On Sat, 9 Nov 2002, Daniele del Re wrote: > > Hi all, > > these are the results fitting the data with the different selections from > Urs: > > http://www.slac.stanford.edu/~daniele/vub/testsele.html > > I fitted the new and the old MC. I also realized that the selection I > used to perform all the tests and scans is a bit different from our newest > selection. I had killTracks=1, killSlowPions=1, doTrackKilling=3 . > > My comments: > > - old-new MC's are the identical as far as the fit result is concerned > - Bch results are stable with all the selections > - B0 results show large differences between g00 and g01,2,3 > - MC shapes: > * B0 g01 - B0 gdan show a very different shape, do we understand > this? > * Bch g02- Bch g03 show a very different shape, do we understand > this? > - data: according to me the three different selections (g01,2,3) > change the data very slightly > > My final comment: > > - Bchs look very good. B0 don't. We should understand which is the cut > that change so much the result from g00 to g01,2,3 for the B0s > > > Daniele > >