Print

---

Hi all,

 I studied the crossfeed B0/Bch and the impact in our analysis.
 We ran on B0 generic MC with Bcha reconstruction and viceversa.

 First of all this is the number of crossfeed reconstructed events
compared with correspondent yields for B0 on B0 MC and Bch on Bch MC:

 http://www.slac.stanford.edu/~daniele/vub/crossfeed.html

 As you see, the fraction of peaking crossfeed is not negligible. This is
not in contraddiction with the Alessio's studies about duplicated events
in different seeds. He showed that the duplicated events were a negligible
effect. In that case the crossfeed was in the breco selected events
itself. Now we are taking into account all B0's and Bch's and the effect
could be different and higher in principle.

 Then I started to look at all recoil variables. I was thinking about the
fact that the cause of the disagreement data MC could be found in this
events.

 In the following html page you find the usual plots for all the crossfeed
events and the comparison data-generic+crossfeed compared with the
data-generic with no crossfeed in.

 http://www.slac.stanford.edu/~daniele/vub/crossfeedcomp.html


 There is also comparison generic with generic+crossfeed.

 http://www.slac.stanford.edu/~daniele/vub/genegenecross.html


 You can see some significant discrepancy in the breco normalized plots,
especially in the depleted sample.

 The systematics on the extraction of BRBR is the following:

 No crossfeed:

 BRBR = 0.017545+-0.00299034

 crossfeed in (in the fitted file):

 BRBR = 0.0164212+-0.003155


 My comments are:

 1) crossfeed is not negligible and should be treated carefully. I think
that Oliver's analysis could be take into account this effect too

 2) including the crossfeed events the comparison data-mc is improved a
    bit

 3) in the final fit the generic+crossfeed MC should be used to get the
    right fraction of bkg events in the semileptonic sample, to get the
    right number of semileptonic events and to correct the Mx shape.

 Daniele