Unexpected behavior after cut in generation of p p > t w z

Asked by Michele Mormile

Hello,

I am using MG5_aMC_v2_8_3_2 to generate events of p p > t w z at NLO. This process is interested by overlap with the resonant p p > t t~ z, so I use the MadSTR plugin to remove it, using either DR and DR with interference.

Since the DR scheme doesn't work really well at high pt, and I need to generate events in that spectrum, I also require a cut in pt on b quarks in the final state, to furthermore reduce the overlap; I only want to keep the soft b quarks which have pt < 30 GeV.

To do that, I apply this cut in SubProcesses/cuts.f :

      do i=1,nexternal ! loop over all external particles
         if (istatus(i).eq.1) then ! final state particle
            if ( abs(ipdg(i)).eq.5) then ! bottom quark
c$$$C apply the pT cut (pT should be less than 30 GeV for the event to
c$$$C pass cuts)
               if ( p(1,i)**2+p(2,i)**2 .gt. 30d0**2 ) then
c$$$C momenta do not pass cuts. Set passcuts_user to false and return
                  passcuts_user=.false.
                  return
               endif
           endif
       endif
    enddo

These are my generation commands in MadSTR:

import model sm-no_b_mass
generate p p > t w- z [QCD]
add process p p > t~ w+ z [QCD]

My problem is that the cut is not really applied, as I have the majority of events with b quarks over the pt threshold, when I look at the pt distributions from the LHE file.

Moreover, I also plotted the distribution of the b quark pt in three cases:
the cut applied to retain the events with b quark pt < 30 GeV (b-veto)
the cut applied to retain the events with b quark pt > 30 GeV (reversed b-veto)
no cut applied (no veto)

What I see is that the reversed b-veto does exactly what it is supposed to do: I have no events with pt < 30 GeV and then I have a fast decrease after 30.
The no veto category has a flat spectrum.
The b-veto category has the same distribution of the no-veto for pt < 30 GeV and then it is very similar to the reversed b-veto for pt > 30 GeV, where there should be no events whatsoever.

Is this behavior correct or am I neglecting something?

Thank you very much for your help

Question information

Language:
English Edit question
Status:
Answered
For:
MadGraph5_aMC@NLO Edit question
Assignee:
marco zaro Edit question
Last query:
Last reply:
Revision history for this message
marco zaro (marco-zaro) said :
#1

Dear Michele,
thank you for writing.
I am puzzled by what you say, as the cuts you implement should exactly do what they are meant for.
When you say that you plot the pt distribution from the event file, you select only the events with a b quark in the final state, don't you?
And this is without any shower?

Let me know

Best,

Marco

Revision history for this message
Michele Mormile (mmor) said :
#2

Dear Marco,

I only select the b quarks in the final state and am doing the plots at LHE level without showering.

Thank you for your help.

Kind regards,
Michele

Revision history for this message
marco zaro (marco-zaro) said :
#3

Hi Michele,
this is strange, let me investigate it a bit more
Cheers,

Marco

> On 12 Nov 2021, at 11:05, Michele Mormile <email address hidden> wrote:
>
> Question #699441 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/699441
>
> Status: Needs information => Open
>
> Michele Mormile gave more information on the question:
> Dear Marco,
>
> I only select the b quarks in the final state and am doing the plots at
> LHE level without showering.
>
> Thank you for your help.
>
> Kind regards,
> Michele
>
> --
> You received this question notification because you are assigned to this
> question.

Revision history for this message
marco zaro (marco-zaro) said :
#4

Dear Michele,
I quote here the answer by Rikkert Frederix

Cuts on the Monte Carlo subtraction terms need to be applied using the n-body kinematics, even if they are contributing to the H-events (which themselves have (n+1)-body kinematics). The reason is that they should subtract the contribution from the Born+Shower from the real-emission. Since the cuts are applied before shower, if the Born passes the cuts, so should the shower subtraction terms.
Hence, I guess that the H-events that pass the cuts and have a b-jet in the final state ones where the real-emission does not pass the cuts, but the MC subtraction terms do.

So in this case, a b jet should be vetoed after shower, like it has been done in sect 4.2 of 2111.03080.
Note that this problem is not there at fixed order

Cheers,

Marco

Revision history for this message
Michele Mormile (mmor) said :
#5

Dear Marco,

I did a further study of this behaviour and found out that almost the entirety of the events with a b quark in the final state have negative event weight, while the few ones with positive weights always respect the cuts.

Is this compatible with your guess?

Cheers,
Michele

Revision history for this message
marco zaro (marco-zaro) said :
#6

Hi Michele,
I would say this makes some sense.
What I strongly suggest, at NLO+PS, is not to apply generation cuts, but to
veto the b-jet after shower
Cheers,

Marco

On Tue, Nov 16, 2021 at 6:15 PM Michele Mormile <
<email address hidden>> wrote:

> Question #699441 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/699441
>
> Michele Mormile posted a new comment:
> Dear Marco,
>
> I did a further study of this behaviour and found out that almost the
> entirety of the events with a b quark in the final state have negative
> event weight, while the few ones with positive weights always respect
> the cuts.
>
> Is this compatible with your guess?
>
> Cheers,
> Michele
>
> --
> You received this question notification because you are assigned to this
> question.
>

Can you help with this problem?

Provide an answer of your own, or ask Michele Mormile for more information if necessary.

To post a message you must log in.