Hi all,

 I studied a bit more in detail the result of the ratio BR(Bch)/BR(B0) on
generic MC.

 With the new statistics (~220 fb-1) I get

 total  BRBR = 0.013032  +- 0.00109235
 B0     BRBR = 0.0133296 +- 0.00181513
 Bch    BRBR = 0.0128752 +- 0.00137274

 ratio = 0.96591 +- 0.167051


 Since a possible problem could be due to crossfeed I performed two
checks:

 - fit the ratio adding a cut on the integrated purity (only in the fitted
sample, not in the models)

 - fit the ratio changing the relative statistics NB0/NBch. If there is a
strong effect of the crossfeed it should show up with this test.

 Here the results

cut on purity
-------------

no cut      ratio = 0.96591 +- 0.167051
intpur>.3   ratio = 0.896821 +- 0.161344
intpur>.4   ratio = 0.901798 +- 0.180097
intpur>.5   ratio = 0.941965 +- 0.19304
intpur>.6   ratio = 0.999017 +- 0.227691
intpur>.7   ratio = 1.35907 +- 0.387477
intpur>.8   ratio = 1.37365 +- 0.548486


different relative B0-Bch statistics
------------------------------------

B0 1/5      ratio = 0.692619 +- 0.24179
B0 1/2      ratio = 1.09417 +- 0.291523
B0 4/5      ratio = 0.995098 +- 0.185956

Bch 1/5     ratio = 1.41696 +- 0.361108
Bch 1/2     ratio = 1.16043 +- 0.225162
Bch 4/5     ratio = 1.00219 +- 0.176967


I don't see any problem in these numbers.

I should be stressed that the statistics is very limited in order to say
that the fit is not biased.

I am also running a test, varying the generated ratio but, again, with
this statistics we cannot say too much.


 Daniele