Fewer LHE events than requested; cannot identify pole (if that is the problem). Process is ttgamma+j.

Hi,

1) How do you handle the double counting between your various (overlapping) processes?
2) You have cut for the collinear part (which might be overwritten if you use MLM for the QCD double counting) but you do not present any cut for the soft divergence.
Do you have some?

Cheers,

Olivier
> On Jan 5, 2017, at 02:16, Ben Nachman <email address hidden> wrote:
>
> New question #429606 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/429606
>
> Hello,
>
> I'm generating ttbar+gamma with up to one extra jet in the matrix element. The radiation from the top decay is non-negligible at low photon pT, we we are doing e.g.
>
> generate p p > t t~ > b b~ u~ d e+ ve a j
>
> However, the number of LHE events is (much) less than the number requested. Reading other questions, it seems like this could be due to integrating over a divergence. However, I'm unable to find the problem. Below are a list of similar processes that I've tried; some work (i.e. number of requested LHE events = number delivered) and some do not:
>
> Works:
> generate u u~ > t t~ > b b~ u~ d e+ ve j a
> generate g g > t t~ > b b~ u~ d e+ ve g
> generate g g > t t~ > b b~ u~ d e+ ve a
>
> Doesn't work:
> g g > t t~ > b b~ u~ d e+ ve j a
> g g > t t~ > b b~ u~ d e+ ve g a
> g g > t t~ > b b~ mu- vm~ e+ ve a g
>
> So the problem is only with the gg initial state and requires both the photon and the gluon in the final state. Any ideas or further checks you have would be most appreciated. Thank you!
>
> Sincerely,
> Ben
>
> PS - the b's are treated as massless, but the process g g > t t~ > b b~ mu- vm~ e+ ve a g still doesn't work even if I add
>
> 0.2 = drjj ! min distance between jets
> 0.2 = drbb ! min distance between b's
> 0.4 = draa ! min distance between gammas
> 0.2 = drbj ! min distance between b and jet
> 0.2 = draj ! min distance between gamma and jet
> 0.4 = drjl ! min distance between jet and lepton
> 0.2 = drab ! min distance between gamma and b
> 0.2 = drbl ! min distance between b and lepton
> 0.2 = dral ! min distance between gamma and lepton
>
> to the run card.
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Hi Olivier,

Thank you for your quick response!

1) CKKW-L (I guess this only applies for the jets and not the photon)

2) Sorry, I did not specify them, but we have cuts for all the objects that could have issues with a soft divergence:

 20 = ptj ! minimum pt for the jets
 20 = ptb ! minimum pt for the b
 10 = pta ! minimum pt for the photons

Sincerely,
Ben

> 1) CKKW-L (I guess this only applies for the jets and not the photon)

Correct, so you indeed have a double counting for the photon.
I’m also surprised that your cut does not seem to fit for any CKKW-L implementation that I’m aware of.
Which type of CKKW-L algorithm (i.e. which scale did you use for the clustering and which variable/value are you using to cut?)

Since you use CKKW-L, you also keep ickkw=0 and xqcut=0 right?

> 2) Sorry, I did not specify them, but we have cuts for all the objects
> that could have issues with a soft divergence:
>
> 20 = ptj ! minimum pt for the jets
> 20 = ptb ! minimum pt for the b
> 10 = pta ! minimum pt for the photons

So a priori all divergency should be cover.
Since you have so many particle in the final state, it might also be a phase-space integration problem.
Did you look at the result for each channel? and spot the channel of integration which miss the targeted number of events?
(you can click on the luminosity title to order them according to the luminosity). The problematic channel should have very small luminosity.

Cheers,

Olivier

> On Jan 5, 2017, at 13:47, Ben Nachman <email address hidden> wrote:
>
> Question #429606 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/429606
>
> Status: Answered => Open
>
> Ben Nachman is still having a problem:
> Hi Olivier,
>
> Thank you for your quick response!
>
> 1) CKKW-L (I guess this only applies for the jets and not the photon)
>
> 2) Sorry, I did not specify them, but we have cuts for all the objects
> that could have issues with a soft divergence:
>
> 20 = ptj ! minimum pt for the jets
> 20 = ptb ! minimum pt for the b
> 10 = pta ! minimum pt for the photons
>
> Sincerely,
> Ben
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

> 1) CKKW-L (I guess this only applies for the jets and not the photon)

Correct, so you indeed have a double counting for the photon.
I’m also surprised that your cut does not seem to fit for any CKKW-L implementation that I’m aware of.
Which type of CKKW-L algorithm (i.e. which scale did you use for the clustering and which variable/value are you using to cut?)

Since you use CKKW-L, you also keep ickkw=0 and xqcut=0 right?

> 2) Sorry, I did not specify them, but we have cuts for all the objects
> that could have issues with a soft divergence:
>
> 20 = ptj ! minimum pt for the jets
> 20 = ptb ! minimum pt for the b
> 10 = pta ! minimum pt for the photons

So a priori all divergency should be cover.
Since you have so many particle in the final state, it might also be a phase-space integration problem.
Did you look at the result for each channel? and spot the channel of integration which miss the targeted number of events?
(you can click on the luminosity title to order them according to the luminosity). The problematic channel should have very small luminosity.

Cheers,

Olivier

> On Jan 5, 2017, at 13:47, Ben Nachman <email address hidden> wrote:
>
> Question #429606 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/429606
>
> Status: Answered => Open
>
> Ben Nachman is still having a problem:
> Hi Olivier,
>
> Thank you for your quick response!
>
> 1) CKKW-L (I guess this only applies for the jets and not the photon)
>
> 2) Sorry, I did not specify them, but we have cuts for all the objects
> that could have issues with a soft divergence:
>
> 20 = ptj ! minimum pt for the jets
> 20 = ptb ! minimum pt for the b
> 10 = pta ! minimum pt for the photons
>
> Sincerely,
> Ben
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Hi Olivier,

Thanks again for your help! Yes, we have ickkw = 0 and xqcut = 0. We are using kT with 30 GeV.

What exactly do you mean by channel? In the original post, you can see that the culprit is the g g > t t~ > b b~ mu- vm~ e+ ve a g and not e.g u u~ > t t~ > b b~ mu- vm~ e+ ve a g. If there is some other useful information I can send you from the output folder, please do let me know!

Sincerely,
Ben

Hi,

you can look at this page
HTML/run_01/results.html
(change run_01 by the appropriate name if needed).
and you will see all the channels, you can then organize them by luminosity.

Cheers,

Olivier

> On Jan 9, 2017, at 07:38, Ben Nachman <email address hidden> wrote:
>
> Question #429606 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/429606
>
> Status: Answered => Open
>
> Ben Nachman is still having a problem:
> Hi Olivier,
>
> Thanks again for your help! Yes, we have ickkw = 0 and xqcut = 0. We
> are using kT with 30 GeV.
>
> What exactly do you mean by channel? In the original post, you can see
> that the culprit is the g g > t t~ > b b~ mu- vm~ e+ ve a g and not e.g
> u u~ > t t~ > b b~ mu- vm~ e+ ve a g. If there is some other useful
> information I can send you from the output folder, please do let me
> know!
>
> Sincerely,
> Ben
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Hi Olivier,

Great, thank you! I see many channels with 0 luminosity. What information can I provide? The graphs have names like "G29.0040". I can also send you the output in zipped format if that would be useful.

Sincerely,
Ben

This question was expired because it remained in the 'Open' state without activity for the last 15 days.

Hello,

Please let me know if I can provide any further information. Thank you for your help!

Sincerely,
Ben

This question was expired because it remained in the 'Open' state without activity for the last 15 days.

Hello,

I know you are very busy - please let me know what information I can provide to help get to the bottom of this issue. Thank you!

Sincerely,
Ben

Hi Ben,

The channel with zero luminosity are problematic only if the cross-section for that channel is different of zero.
So I guess you have to look at another one when you have the channel, then you can take a look at the following link:
https://answers.launchpad.net/mg5amcnlo/+question/187387
to see which topology of diagram is responsible.

Cheers,

Olivier