Lepton efficencies in CMS card

[Philipp]

Hello,

the default CMS card in 3.1.2 features lepton efficiencies that appear quite conservative, especially when comparing to e.g. the ones reported in the Higgs papers.

I'd like to understand why that is the case.
A) What is the source (detector perforce report or a specific analysis) for the efficiencies used in the card?
B) Who is the author of the card that might know more, if its not handled by the core Delphes team?
c) Why is there a split between tracking and reconstruction efficiencies? As far as I know the differentiation is not performed by the experiments and the tracking efficiency should be floe to unity?

Thanks for any hints regarding the above questions.

Cheers
Philipp

[Michele Selvaggi]

Hi Philipp,

the values of leptons and photons efficiencies are not tuned to any particular analysis.
The reason for this is that every analysis has its own object definition. The total efficiency for a given analysis depends on several things
such as trigger, reconstruction and so on.

So to answer more specifically to your questions:

1) the values in the cards are purely indicative (unlike the tracking efficiencies). You are supposed to adjust them according to your specific analysis.

2) see above.

3) We have decided to split them because the tracking efficiencies will affect the Particle Flow algorithm and the lambda user is not supposed to modify them. On the other hand the "residual" (this is how you should think about them, rather than "reconstruction") efficiencies should be modified by the user to account for trigger, reconstruction or any residual effect not taken into account by the tracking.

So the prescription is something like:

• obtain from a similar published analysis some parametrisation of the total efficiency eff_tot
• apply in Delphes an efficiency eff_tot/eff_trk where eff_trk is the one given by delphes.

If you do not have time to produce the full eff_tot/eff_trk parametrisation just apply an constant factor so that eff_trk*eff ~eff_tot.

Does it make sense?

Michele

[Philipp]

Hi Michele,

thanks for your almost instantaneous reply! My questions are answered.

Philipp