hi Pedro,

I do not understand what you are doing and what point you are trying to make. This is the situation as I see it in the context of the ttbar dilepton analysis.

(1) We have known for a long time that muons can fake electrons.

(2) We also know that electrons do not fake muons (to a large degree)

(3) Thus if an electron and a muon share the same CTF track, the electron should be rejected but the muon kept (provided it passes the cuts)

Operationally, in part (3) instead of checking that the CTF tracks are the same, we have been applying a deltaR<0.1 cut. This is a remnant of the times (CMSSW 16x) where it was not easy to find the CTF track associated with the electron.

The deltaR cut is done using the p4 of the e/mu because the direction of the e/mu in both cases comes from tracking, so this is the most natural way to do a match. You do not get exactly deltaR=0 because - the electron uses the GSF track + a beamspot correction (**) - the muon uses the CTF track in most cases, but occasionally the global track. A cut on deltaR<0.1 does the job pretty well. I agree that a cut using the ecal position would work just as well.

The only question that remains is: Do we loose W->e electrons in dilepton events with this procedure. There are three ways that this can happen

(a) A ttbar->e-mu event where the electron and muon happen to be within deltaR<0.1 in the tracker or within whatever deltaR cut you propose in the ecal. I call this "fratricide". Actually, there is a much more significant problem related to the isolation cut. Every time the electron and the muon are within deltaR of 0.4 they will both fail the isolation cut. So, if you want to recover this small inefficiency you should first go after the deltaR~0.4 fratricide isolation problem.

(b) The electron randomly overlaps with one uncorrelated segment in the muon chambers, therefore the electron track makes a (crappy) tracker muon. We looked at that some time ago, and it was not a big deal, but it could be useful to look at it again. One can imagine putting some quality cuts on the muon.

(c) There is a muon (not from W->mu) within deltaR of 0.1. However, if this is the case the electron will be rejected because of the track isolation requirement (unless the muon is very low pt AND there is not much else around it). Whether this event gets rejected by the deltaR<0.1 cut or the track isolation cut, the end result is the same, so I don't really care.

The bottom line to me is:

- The deltaR(e-mu)<0.1 requirement is fine. - It would be more elegant and precise to compare CTF tracks, but the end result will be the the same unless the fratricide issue in the isolation is addressed (*). I say "almost" because of the very low pt muon case in part (c). In this case the CTF track match is better. - It would be a good idea to see whether we should put some tighter muon requirement on the muons that are used in this veto (whether it is done with deltaR-objects, deltaR-ecal, or CTF track matches).

Claudio

(**) https://hypernews.cern.ch/HyperNews/CMS/get/egamma/673/1.html (*) I am not advocating that we try to fix the fratricide problem now. I think it is a small effect, it is be well modeled by MC, and it would be a big pain in the neck to implement correctly (we have better things to do)

-- MaydaVelasco - 09-May-2010