Matching and jet cuts

Hello Daniele,

First there is an issue with your process definition in your mail

generate p p > n1 x1-, x1-> n1 e- ve~ @1
add process p p > n1 x1-, x1-> n1 e- ve~ @2

I guess this is just a typo right?

Now coming to your question:

in case your have T = auto_ptj_mjj then ptj and mjj are set to the value of xqcut, that is what actually defines the kt distance between the partons, and between the partons and the beam line. Then Pythia will "fill" the phase-space with soft jets (not resolvable in the sense of the matching scale (qcut). The cross-section you get after the pythia steps correspond actually to the one after the matching, which means something close to the inclusive case where no cut on ptj is applied.

Concerning the cuts that are applied, as said before ptj and mjj are most probably set to xqcut (hence no effect on the physics output), but any Ht cut or eta cut you would decide to apply might have an effect.

Hope this helps

Simon

I realize I wrote some of the explanations in a non very precise way; let me rephrase/precise one or two things.

a) The problem with your process is that in the add process line , there is no additional jet.
b) about the soft jets from Pythia, extra-partons with a kT harder than the matching scale are allowed for the highest multiplicity sample.
c)"which means something close to the inclusive case where no cut on ptj is applied." : by inclusive I meant the O-parton multiplicity, inclusive. So to put in another way: xsec(X+0,1 partons after the jet matching)~xsec(0 parton) .

Simon

Hi Simon, thanks for the answer.

Yes there was a typo in the second process, the correct is

generate p p > n1 x1-, x1-> n1 e- ve~ @1
add process p p > n1 x1- j, x1-> n1 e- ve~ @2

Regarding your first answer I try to explain if I understood correct:

before matching (before pythia) the cuts on the ptj are set equal to the xqcut variable, so that mean that I will not have, for example, jet with pt<30 GeV.
Then pythia will fill the phase space with also soft jets, so it will put jets with pt<30 GeV and then matching procedure between the two is performed.

But in this final cross section I will not have any cut on the ptj, so I can have also really soft jet (added by pythia) that contributes to my cs if I understood correctly. In other words the xqcut will cut ptj before pythia, but we will loose this cut after pythia, having at the end events with all kind of ptj, is that correct?
So in case I want to apply some cuts on the jets after matching do I have to do it by hand from the root/lhe files?

Regarding point c) of your second answer, does that mean that the cross section at parton level at 0 jet should be equal to the cross section 0+1 jet matched within the 20% that is usually said on the MG documentation?

Hi Daniele,

"before matching (before pythia) the cuts on the ptj are set equal to the xqcut variable, so that mean that I will not have, for example, jet with pt<30 GeV.
Then pythia will fill the phase space with also soft jets, so it will put jets with pt<30 GeV and then matching procedure between the two is performed."

Well, this is automatic, the showerkt just tests the presence of hard emission from PS, and compare either with the qcut or with the softest ME partons. From that it decides to veto or not the event.

"But in this final cross section I will not have any cut on the ptj, so I can have also really soft jet (added by pythia) that contributes to my cs if I understood correctly. In other words the xqcut will cut ptj before pythia, but we will loose this cut after pythia, having at the end events with all kind of ptj, is that correct?"

Yes, xqcut acts ONLY in the ME computation (=>LHE output), and the Qcut makes sure that ME calculation takes care about what is hard, PS about what is soft, and with no overlap between the two regions.

"Regarding point c) of your second answer, does that mean that the cross section at parton level at 0 jet should be equal to the cross section 0+1 jet matched within the 20% that is usually said on the MG documentation?"

Yes, since the jet matching 'replaces' the leading jets from the showering by jets from the matrix-element calculation, you therefore expect to find a 'matched' xsec that is close to the LO 'unvetoed' xsec.

Thanks a lot again, just a final remark about the % different between the two cross section to see if it is really clear to me

> The cross-section you get after the pythia steps correspond actually to the one after the matching, which means something close to the inclusive case where no cut on ptj is applied

> xsec(X+0,1 partons after the jet matching)~xsec(0 parton)

> expect to find a 'matched' xsec that is close to the LO 'unvetoed' xsec

The inclusive case in my case is only given by

generate p p > n1 x1-, x1-> n1 e- ve~ @1

so when you say "no cut on ptj is applied" or "unvetoed" looks strange to me, since in this case I don't have any jet in my matrix element processes and so don't make sense to apply a ptj cut... is that correct?

and a final question:

so if I want to apply some kind of cuts on a matched sample do I have to do it by hand from the pythia output files?

Thanks in advance again and sorry for the trivial questions... I always get confused by matching procedure...

Thanks for your answers!

Thanks de Visscher, that solved my question.