The rejection of jets faking a certain particle hypothesis (photons, electrons) is defined as the ratio between true jets within a certain acceptance and the number of reconstructed fake candidates in the same jet sample

R = N(true jets)/N(fake candidates)

This can also be rewritten as

R = N(events)*F/N(fake candidates)

where F is the average number of jets (within acceptance) per event. Typically, in order to have larger statistics, one filter events; if eff is the event efficiency of the filter, then

R = N(events after filter)/eff * F/N(fake candidates) = N(events after filter)/N(fake candidates) * F/eff

where I have rearranged the order of the factors in order to isolate a part, F/eff, that needs to be obtained by looking at unfiltered events, and a second part, N(events after filter)/N(fake candidates), that can be evaluated just looking at ntuples obtained from reconstruction of filtered events.

One final comment: the jets acceptance should match the acceptance criteria used to identify the particles under study (photons, electrons, ..) and used when evaluating also signal efficiencies. For instance, one may require 0<|eta|<1.37 or 1.52<|eta|<2.37 and ET>25 GeV for photon efficiency and fake photon jet rejection.

-- GiovanniMarchiori - 2009-09-30