Hi all-

I am making a mode by mode study of crossfeed on the breco side. My plan is to 
check for B0 <--> ChB xfeed per mode and then further check for B+ <--> B- 
xfeed or B0<-->B0bar xfeed if there is no B0<-->ChB xfeed for all of the 
breco modes. 
This part of the study is more or less straight forward. Some care needs to be 
taken to pick up the right B from the MC truth block however. The existing 
truth method in the Vub code looks for the recoiling Vxb decay and assign the 
breco B to be the other B in the MC block. If you can't identify a Vxb decay 
in the mcTruth() method then you have to take care to pick the right B. But, 
if you have a Vxb decay on one side, there is no confusion about which B to 
pick. 
Past this, I'm planning to go into detail to and loop through the MC block to 
look for instances where we have the wrong D seed. Care has to be taken in 
picking the right D in the case that there are more than one D as a daughter 
of the B. 
I'll also look for cases when gain/lose pi,pi0,K,Ks while making sure to split 
up the D seed and the B decay parts separately in the decay
For example, even though they are different decay modes, B->D0 pi, D0->Kpi pi0 
would look identical to B->D0 pi pi0, D0->K pi if you looked at the D seed 
particle and counted the pi/pi0/K/Ks in the event without separating the D 
seed part of the decay away from the B mode part.

What I've done so far is to look for B0 <--> B+ xfeed and then B+ <--> B- and 
B0 <--> B0bar xfeed per mode. 
The crossfeed is defined as (total signal events – total correct signal 
events)/(total signal events). The error on the crossfeed shouldn't be taken 
too seriously, I just filled it with something simple for now that doesn't 
hold for low statistics/poor fits. It might be easier if I fit the number of 
incorrect events to mes rather than taking the differece of total and 
correct.  This could also treat the cases where the fits for the number of 
correct events return more signal events than the fits for the total number 
of events.
The number of signal events come from Mes fits on the entire generic sample on 
events with a semileptonic decay identified

See http://www.slac.stanford.edu/~ejhill/Internal/Vub/20031028/out1.html for a 
table of the results.  There are no cuts applied (other than it being paired 
to a semileptonic event)

Cheers,
Ed