 |
 |
Hi, some updates. First of all, new tags (cb-060801) of both VVF and BAD/note1383 are available. I fixed a couple of things in the macro which performs the S/P study, it should be now straightforward to get the numbers out (it even tells you in which cells of the spreadsheet you should paste them!). I think it would be nice to see what changes when applying tighter Breco purity cuts. Once the S/P factors are determined, one can put them in a file which is supplied to VVF with the -fixSBratio flag. The S/P systematic studies can be performed by using the VVF flag -SPseed, have a look in job_SP.pl on how to use it. > * Concezio has provided a documentation in BAD1383 (Appendix E) > He will correct it a bit since several captions do not coincide with > the quantities shown. > I have commited the changes into CVS and tagged BAD/note1383. The numbers now include the latest results with the yields required to be positive. I am currently running the 100 randomized fits in order to see what is the systematic effect due to our imperfect knowledge of the S/P ratios. > * We have observed that - when applying the correction factor - that > yields in the mES can become negative (since the correction factor > can become negative when fitting the correction factor as a function > of mX with e.g. a second-order polynomial) > > This leads to the question if we want to constrain in general the > (signal) yields > from the mES fit being positive. > We had already a discussion at Montreal about this. I do not remember > if we decided there to allow the yields to become negative in order > to avoid a possible bias. Probably, such a bias is not an issue and > we should enforce the mES yield to be positive? > I performed this exercise, the S/P ratio does not change much in the significant bins. It changes somewhat in the low statistics bins, but the effect on the PBRBR is negligible (it affects the Vub yield on the third significant digit at most). > * Looking into the mES fit on the data-depleted sample which is important > input to our correction factor we observe that many fits have problem in > the region around the anchor point of the Peaking BG PDF. > > I think we should consider trying to apply the mES correction and see > if the fits behave in a better way. Of course, this does not correctly > correct for the endpoint but is maybe not so important. > Besides studying higher purity cuts: Chukwudi and Wolfgang, could you > please try this? (Wolfgang did already something in this direction.) > > Cheers, > Heiko > > P.S.: I will leave next Saturday if there is any need for discussion. > Concezio is not any more reachable. I will probably check my emails around mid August. Ciao, Concezio.