Hi Vera,

> I presume the two fits listed are the same fits as listed in Table 11,
> just translating the fit results into Rc rather than Ru.  Is this what
was done?
>

Yes, it is.

> In looking at the fits, I do not understand the errors
>   a) compared to Table 11

let's take the Bch case.

Ru has 86 signal events with 85 background events (for Mx< 1.55).
The error from the mes fit is 16.
The error from the fit is 17.

R_c has ~800 events (full range in Mx).
The error from the mes fit (adding in quadrature all errors in each
single bin) is 36
The error from the fit is 50.

I studied this apparent diagreement.

Fitting Rc fixing the remaining two components (R_u and R_o) I get
37 as error from the fit.

Then the disagreement is due the fact that three componenents are floated
at the same time. BTW fit on R_u is under control.

>   b) for Rc from the depleted sample and the various subsamples;
>      the weighted average of the subsamples are not consistent with the
>      results from the total sample and its error

This disagreemet was due to a wrong N(Bch)/N(B0) reweighting. I remind you
that the MC is reweighted in order to get the same N(Bch)/N(B0) ratio as
in data.
In this R_c fit I was just using the reweighting for the Vub MC and not
for the Vcb (generic) MC. This implied a bias in the full sample. I
fixd it. This is the last set of numbers I have.

Enriched sample:

All  BRBR = 0.976831 +- 0.057952(stat)

B0   BRBR = 0.967336 +- 0.105336(stat)
Bch  BRBR = 0.986532 +- 0.0677618(stat)
ele  BRBR = 0.952345 +- 0.072551(stat)
mu   BRBR = 0.983558 +- 0.0897417(stat)
run1 BRBR = 1.07744 +- 0.0999208(stat)
run2 BRBR = 0.9326 +- 0.0637472(stat)
sb1  BRBR = 1.03512 +- 0.105587(stat)
sb2  BRBR = 1.02624 +- 0.0781438(stat)
sb3  BRBR = 0.904608 +- 0.0749679(stat)

Depleted sample:

All  BRBR = 0.998844 +- 0.0296921(stat)

B0   BRBR = 0.980276 +- 0.0584093(stat)
Bch  BRBR = 1.00257 +- 0.0341011(stat)
ele  BRBR = 1.00265 +- 0.0395835(stat)
mu   BRBR = 0.976236 +- 0.0404054(stat)
run1 BRBR = 1.07242 +- 0.0469847(stat)
run2 BRBR = 0.969857 +- 0.0325255(stat)
sb1  BRBR = 0.987284 +- 0.0614623(stat)
sb2  BRBR = 1.04291 +- 0.0412906(stat)
sb3  BRBR = 0.962242 +- 0.0402913(stat)


Now this set of numbers is more coherent. A bit of disagreement is still
in the electrons-muons mean but this could be related to the fact that the
two samples are not uncorrelated.

>   c) the "other" background is taken from MC,
>      how different is C_o in the enhanced and depleted samples?

they are comptible within the stat error

enriched :  C_o = 0.405776 +- 0.290029

depleted :  C_o = 0.28122 +- 0.163354

>   d) when you fit the depleted sample, did you fix C_u?

no, I don't. Actually in these fits C_u, C_c and C_o are floated.

For the remining issues Urs is producing plots.

Cheers,

Daniele