JetDeltaPt from examples/Example3.C

[Dongsub Lee]

Dear Delphes team,

Hi.
I checked the JetDeltaPt which is normalized(by pT of a jet) difference between pT of summation of all constituents of jet and a pT of jet itself is not centered at 0 and have quite large discrepancy, around 15% for simple p p > j j samples which is generated by MG5 and showered by Pythia8.

I saw that discrepancy by my own code first, and I also saw the same thing happens with examples/Example3.C.

I simply ran Pythia8 and Delphes in pipe line of MG5 with default Delphes card(delphes_card_ATALS.dat) after replacing input array of jet finder to EFlowMerger/eflow from Calorimeter/towers.

Currently I'm using Delphes v3.4.1, and I addressed all branches of delphes_events.root in my code.

I've checked that whether I correctly accessed all constituents of a jet or not, but I think that I did not make that kind of mistake.

So here is my question.

1.
If it is not a usual situation, then did I missed some constituents of a jet?
I mean, is there any constituents which is not included in the TRefArray of Jet->Constituents but clustered as a jet by FastJet?

2.
If not, so that it is usual case, then is it okay to use such jets having pT discrepancy with its constituents?
Can I assure that there is nothing I missed?

[Dongsub Lee]

Result from DELPHES/examples/Example3.C

[Dongsub Lee]

Delphes card

[Dongsub Lee]

I uploaded result plot and Delphes card used in simulation.

Because of the size limit of file I couldn't upload a sample, but you can easily(and quickly; it would not take long time) reproduce sample with MadGraph5.
I used sqrt(s) = 14TeV and didn't change anything from MadGraph5 configuration.
Event process is [ p p > j j ]

[Michele Selvaggi]

Most likely it is because a jet energy scale is applied:

​https://github.com/delphes/delphes/blob/master/cards/delphes_card_ATLAS.tcl#L630-L640

can try putting the value to 1 in the ScaleFormula ?

That is happens simply because the Jet momentum is rescaled but not the jet constituents.