phylsix / firehydrant Goto Github PK

denominator: number of all simulated signal events passing gen filters,
numerator: number of events where two darkphotons are matched by reco-ed leptonJets, leading and subleading leptonJets are matched with gen darkphotons.

get those numbers for all mass points, decay lengths, final states combinations. Color them in a 2D map with mass of boundstates and darkphoton as XY axes, each decay length and final state goes into an individual map.

The default leptonJet pair selection method , leading and subleading, giving us a unphysical tail in the low end in the invariant mass spectrum. I suspect that in those events, the second darkphotons are actually NOT reconstructed, the subleading selection picks up some fake leptonJets and paired up with the correctly reconstructed leptonJets.

To confirm this, we can first filter out events where both darkphotons are matched with reconstructed leptonJets, and then count the fraction of events where both leading and subleading leptonJets matched with gen darkphotons. If these numbers are close to one, then I believe selection of the leading two is enough for our purpose. And dphi>2.5 shall kill most wrong combinations.

should incorporate:

• genwgt
• pileup reweighting

• (nlo weights)