Hi all,
I repeated the comparison SemiExcl-Ibu running the run1 Dstar reduced
collections. Compare
http://www.slac.stanford.edu/BFROOT/www/Physics/Analysis/AWG/EHBDOC/SemiExcl/
dstarurs/mainB0Dstar.html
with
http://www.slac.stanford.edu/~daniele/vub/testIbu/sx-2.html
and the legenda for the mode is, for instance, in:
~daniele/devel/testCP/CPFramework/tagB.F
The overall yield has a loss of around 10%. Looking more in
detail I see (modes with a yield > 50 events):
- D*pi ~10% less (all D modes)
- D*pipi0 ~5% less (~all D modes)
- D*pi2pi0 ~compatible
- D*3pi ~10% less (all except D0->kpip0)
- D*omegah ~10% more (except D-k3pi) purity higher
- D*D* compatible
- D*3pi2pi0 compatible
For the remaining modes:
- mode 17 (Dstark2pi0ks_1.8-2.2GeV) is gone
- mode 18 not present in Cpf, some events in Ibu (maybe a flip with the
previous one?)
- mode 21 (Dstar2kpi2pi0_Ds*) factor 2-3 more in Ibu
- mode 22 (Dstar5pi) not present in semiexcl!!!! I found a bug in the
table, 0 yield for this mode. In Ibu just background but this does not
make sense since the table says that the yield is 0. I discovered that
modes with 0 yield have random value of intergrated purity. WE HAVE TO FIX
THIS.
- mode 24 (Dstar2K3pi_<2.7GeV) cleaner in Ibu
- mode 30 is not present in Ibu
- mode 43 (B -> D 2k pi pi0) not present in Ibu.
- mode 47 (B -> D 2k pi 2pi0...) is not present in Ibu
- mode 53 not present in Ibu
I will try to understand if the some of these discrepancies are
time(subsample)-dependent.
Daniele
On Fri, 28 Dec 2001, Daniele del Re wrote:
>
> Hi all,
>
> I performed a comparison ibu vs cpf looking at mes plot mode by mode.
>
> I started with the tags used by Urs for the mc-data production (see
> IslBrecoilUser/tags-112101). In ran just one reduced superblock1
> collection for the D* (~2fb-1). The statisics is low but the results seem
> to be compatible. The only relevant difference is in the B mode 38-39 (B
> -> D 2k pi in K* window). There was a bug in the code for this mode. Now it is
> fixed by Riccardo in the new tags of CompositionTools (rf-121701).
>
> I repeated the study with the new tags (IslBrecoilUser/tags-122101). We
> have to add some new tags to make things working fine in analysis-12a. You
> have to take the tags in
>
> http://www.slac.stanford.edu/BFROOT/www/Physics/CP/beta/doc/postCP/tags-anal12
>
> I ran on two colletions (one of superblock1 and one of superblock2).
> Again the results are very compatible. Just problems in some modes with
> low yields:
>
> - mode 53 seems to be gone
> - some disagreement in mode with high multiplicity and with many kaons
> (19-20-35-45)
>
> No other evident bugs were found. More studies should be needed but since
> these modes are not crucial we can post-pone them.
>
>
> Then the CONCLUSIONS are:
>
> - good agreement but I need to run on the full sample of reduced
> collections to say more.
> - no lack of yield in the main modes
> - Dpi mode is fine.
> - We have to move to the new tags with bug fixes for the new production
>
> - I don't understand numbers presented by Urs in
>
> http://www.slac.stanford.edu/BFROOT/www/Physics/Analysis/AWG/InclusiveSL/islreco
> il/modes/mesallevents-0-1.eps
>
> All D*pi have a lack of ~25% in yields. In my study I did not see any
> problem in this mode.
> I tried to look at the production root files for the D* and I noticed
> something strange. For instance in:
>
> /nfs/farm/babar/AWG/ISL/recoil/root/breco-112101/data/
> dstar_092701_superblock1_2000_b2-s3_aa.root
>
> I see ~630 events (originally the collection had 2000 events) and the run
> range is 15300-15600. Runnning on the same collection I got ~1750 events
> with a run range of 15300-16400. Something is wrong here and could explain
> the difference in yield that we are observing. Urs, may you run on this
> collection and check. Thanks,
>
> Daniele
>
>
>
>
> On Fri, 21 Dec 2001, Urs Langenegger wrote:
>
> >
> > Hoi,
> >
> > ok, I redid the mode-dependent study, you can find them starting at
> > http://www.slac.stanford.edu/BFROOT/www/Physics/Analysis/AWG/InclusiveSL/islrecoil/modes/modes.html
> >
> > The signal pdf is a gaussian for the mode-dependent mes plots. All
> > modes I see are shown. No cut on purity has been applied. All data in
> > Run 1 in superblock 1 and 2 are used.
> >
> > I now see 41376 B candidates, fitted with CB, this is due to the
> > patched B+ modes, see
> >
> > http://www.slac.stanford.edu/BFROOT/www/Physics/Analysis/AWG/InclusiveSL/islrecoil/modes/mesallevents-0-1.eps
> >
> > You can run the relevant macro by doing
> >
> > cd ~ursl/root/anaRecoil
> > root
> > .L .../recoil.C // wherever you have your IBU stuff
> > loadMergedFiles()
> > plotMes(2) // the argument is the offset for the file
> >
> > Note: Please edit recoil.C first, so that you dump the plots into
> > another location (search for 'name' in plotMes() ...). Of course, you
> > can run it on your own rootfiles as well, just edit loadMergedFiles().
> >
> >
> > Merry Christmas, now for good!
> > --U.
> >
>
>
>
|