Hi all,
I made a first check of the analysis code committed in IslBrecoilUser.
I perfomed the test on B0 cocktail MC comparing with results obtained
using SemiExclUser+CPFramework.
For IslBrecoilUser I used:
- root file produced by the last version of IbuNtpDump.cc (and tcls).
They are the ones in Urs page.
- the recoilNtp.cc and other root stuff in IslBRecoilUser (HEAD)
For SemiExcl + CPFramework side:
- hbook produced by the last version of SemiExclAnalysis.cc
- CPFramework and my routine CpfRecoilStudy.F (basically the source code,
root macros are pretty similar to it)
I compared the histograms implemented by now, some MC plots and reco
plots after lepton and Q=0 requests.
MC :
----
- p*, Mx are ok
- some disagreement in the number of semileptonic event (b->clnu). The
Isl... code seems to have less event (you expect ~21% of the total)
http://www.slac.stanford.edu/~daniele/vub/nVcb_root.ps
http://www.slac.stanford.edu/~daniele/vub/nVcb_cpf.ps
(I ran on different statistics)
RECO:
-----
(cpf was run on ~5% more events)
- Mx: pretty similar but some shift up in the mean. It seems that there
are less events at Dc/D0 and D* masses.
http://www.slac.stanford.edu/~daniele/vub/Mx_root.ps
http://www.slac.stanford.edu/~daniele/vub/Mx_cpf.ps
- Mv**2 : fine
http://www.slac.stanford.edu/~daniele/vub/Mnu_root.ps
http://www.slac.stanford.edu/~daniele/vub/Mnu_cpf.ps
- Reso Mx: a bump around 3 Gev in Isl...
http://www.slac.stanford.edu/~daniele/vub/resoMnu_root.ps
http://www.slac.stanford.edu/~daniele/vub/resoMnu_cpf.ps
The main disagreement is in the number of b->clnu events in the MC
truth and the Mx resolution.
These effects need to be investigated.
Daniele
|