Our use-case in H->TauTau is for the mu->tau embedded samples for the binomial bin-by-bins we add in addition. This accounts for the uncertainty, which comes from the limited number of trials we use to filter certain kinematics of the tau decays.
It gets relevant for 'emu' channel because of the very large statistics in mu->tau embedded events on one hand and the very small mean value of the fraction of events passing the filter
Hi Andrew,
Thanks for your review!
Our use-case in H->TauTau is for the mu->tau embedded samples for the binomial bin-by-bins we add in addition. This accounts for the uncertainty, which comes from the limited number of trials we use to filter certain kinematics of the tau decays.
Here the current implementation of it:
https://github.com/KIT-CMS/SMRun2Legacy/blob/master/interface/BinomialBinByBin.h https://github.com/KIT-CMS/SMRun2Legacy/blob/master/src/BinomialBinByBin.cc
and the usage in our Morphing for SM HTT ML-based:
https://github.com/KIT-CMS/SMRun2Legacy/blob/master/bin/MorphingSMRun2Legacy.cpp#L578-L585
It gets relevant for 'emu' channel because of the very large statistics in mu->tau embedded events on one hand and the very small mean value of the fraction of events passing the filter
Cheers,
Artur