 |
 |
Hi Francesca, this is my interpretation. We know from Chukwudi's purity study that S/P goes up with tighter cuts on purity. And we have argued to go to Thorsten's model because of problems in our signal yield. So, the original fit model is expected to have a problem at smaller purity cuts and it is expected to have less problems at higher purity cuts as it is likely that the Crystal Ball is able to take into account the tails from the peaking BG. I do not interprete these numbers as a proof though because what is missing is the study on MC which is necessary to see that the bias with the old model becomes resonably small with the original model. In addition, we need to check all MES fits by eye. Cheers, Heiko On Wed, 13 Sep 2006, Francesca Di Lodovico wrote: > Hi Heiko, > > should we then interpret the fact that the partial BRBR is stable at > larger purity cuts due to the fact that the peaking contribution > is smaller? The original fitting model neglects that > component, so when that background is larger the fit might not give the > right result? > > Cheers, > Francesca > > PS I read all the presentations and I do not have anything to add > to what already said. But in particular I think it is really needed to > have a slide on the reweighting in Antonio's presentation. > PPS I'll be on a plane this afternoon, so unfortunately I will not able to > attend the meeting... > > On Tue, 12 Sep 2006, Heiko Lacker wrote: > > > Hi, > > > > I have copied a ppt- and pdf-version of a transparency describing what > > Wolfgang has done concerning the fallback solution: > > http://iktp.tu-dresden.de/~lacker/fallbacksolution.pdf > > http://iktp.tu-dresden.de/~lacker/fallbacksolution.ppt > > > > From what I have seen in Wolfgang's directories it looks like that the > > chi^2 values are good: > > http://www.slac.stanford.edu/~menges/Vub/scra/intpur > > > > Maybe somebody else can also have a look...? > > > > Cheers, > > Heiko > > > > >