Hello folks,
I have been checking the ~10% events that do not get selected by Ibu.
The majour source of losses (~7%) is due to the DeltaE cut. Since the
deltaE definition is slightly different in CPF and Ibu ( the B candidate
is vertexed in one case and not in the other) events can slightly move
around. When we will have the skims this will not be an issue because the
definition of the variables will be consistent. For the time being ,
running on reduced collections, we need to blow up the windows.
I have implemented this in
SemiExclUser rf-051202
CompositionTools rf-051202
BetaCoreTools rf-010402
the remaining 3% is due to the fact that we select the best D* before
doing a B instead of after doing it (as it was in CPF). I do not think we
can do much here
ciao
ric
|