Dear all,
I’m working on a simultaneous fit, and I encountered some normalization problems.
I have different channels with different branching fractions, and they all decay into the same final state. The fit result is a combined amount of signal events, because each individual amount of signal events is fixed to the global one. The background shape is both fitted to the signal region (left histogram), and a sideband region (right histogram)
As soon as I have the fit result, the expected number of signal events for every single channel is calculated, and a signal distribution is drawn into each signal histogram (and the same for the background plots).
Each histogram is plotted with a command similar to this:
RooPlot* frame11267 = BdeltaE.frame() ;
combData.plotOn(frame11267,Cut("Mode==Mode::S11267"),Binning(50),DataError(RooAbsData::SumW2)) ;
simPdf.plotOn(frame11267,Slice(Mode,"S11267"),ProjWData(Mode,combData)) ;
simPdf.plotOn(frame11267,Slice(Mode,"S11267"),Components("untergrund11267"),ProjWData(Mode,combData),LineStyle(kDashed));
All of this seems to work as it should, but it fails at some histograms where no candidates are left in the histogram after the final selection. I think its a normalization problem, I have tried to set a fixed normalization, but it didn’t work. I’ve also tried to set the axis range to really high values, but the blue line is still drawn above the histogram. How can this be fixed?
Thanks a lot,
Sebastian
