Hi, I have been exploring the follwing working hypothesis (which is what Daniele has shown the results with): use all the no Ks bug MC, both the generic and the cocktail. The difference between cocktail and generic is probably small compared with the effect of the ks bug: see for istance http://www.slac.stanford.edu/~rfaccini/phys/vub/noksbug/comparison7mm2allcuts-cockgen.ps (this is done on all the available MC with no bug) We still have to do some work in equalizing the B0/Bch content ... let us discuss this line of attack at the meeting. ciao ric