MadGraph5_aMC@NLO

divergence and cut

Asked by joholi

Hi,

 I generate p p > z j, (z > ve ve~) at the parton level, and change the default run_card with
"0 = ptj" and "0.001 = drjj". I want to generate 50000 events, but get 900 events finally, and the cross section is 5.22e+04 pb. There should be the divergence. Therefore, I take "10 = ptj" and "0.4 = drjj", generate 50000 p p > z j, (z > ve ve~) again, and get 50000 events finally and cross section is 1711 pb.

Does the second results still include the contributions of divergence? In fact, I only take p p > z j, (z > ve ve~) as an example, and my real problem is: How to determine whether the final results include the contributions of divergence for a given process (in addition to the number of events)? Thus, I can make the minimal cut in the run_card and get the physical result without prejudice.

Best,

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Solved by:
Olivier Mattelaer
Solved:
Last query:
Last reply:
Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#1

Hi,

> Does the second results still include the contributions of divergence?

You still have the same dependencies with the 1/q^4 behaviour. Now you decide to cut at some point or another is in itself arbitrary.

> How to determine whether the final results include the contributions of divergence for a given process (in addition to the number of events)?

The real question is how to determine which part of the Phase-space is correctly describe at LO, and which one is correctly describe at NLO, from those which are correctly describe by the parton shower. This question is actually kind of a solve problem if you consider LO/NLO matching merging (like MLM, CKKW, CKKWL, FxFx, UnLOPS,…) I suggest that you read those paper.

Cheers,

Olivier

On 05 Jun 2015, at 13:46, joholi <email address hidden> wrote:

> New question #267805 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/267805
>
> Hi,
>
> I generate p p > z j, (z > ve ve~) at the parton level, and change the default run_card with
> "0 = ptj" and "0.001 = drjj". I want to generate 50000 events, but get 900 events finally, and the cross section is 5.22e+04 pb. There should be the divergence. Therefore, I take "10 = ptj" and "0.4 = drjj", generate 50000 p p > z j, (z > ve ve~) again, and get 50000 events finally and cross section is 1711 pb.
>
> Does the second results still include the contributions of divergence? In fact, I only take p p > z j, (z > ve ve~) as an example, and my real problem is: How to determine whether the final results include the contributions of divergence for a given process (in addition to the number of events)? Thus, I can make the minimal cut in the run_card and get the physical result without prejudice.
>
> Best,
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Revision history for this message
joholi (joholiphy) said : 		#2

Dear Oliver,

The matching is indeed a valid solution for the process with jet.

However, are there some valid solutions for the photon process, such as p p > z \gamma (z > ve ve~). How to determine whether the cut is proper? If there is divergence , the cross section is very sensitive to cut, and further produces the great effect on the normalized histogram.

Best

Revision history for this message
Best Olivier Mattelaer (olivier-mattelaer) said : 		#3

Hi,

> The matching is indeed a valid solution for the process with jet.

In principle the same should be done for photon with photon shower.

> If there is divergence , the cross section is very sensitive to
> cut, and further produces the great effect on the normalized histogram.

The only point is that the cut that you apply at MG level will not impact the
cut that you will apply at the analysis level. So in principle it should not matter.

Cheers,

Olivier

On 05 Jun 2015, at 16:56, joholi <email address hidden> wrote:

> Question #267805 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/267805
>
> Status: Answered => Open
>
> joholi is still having a problem:
> Dear Oliver,
>
> The matching is indeed a valid solution for the process with jet.
>
> However, are there some valid solutions for the photon process, such as
> p p > z \gamma (z > ve ve~). How to determine whether the cut is
> proper? If there is divergence , the cross section is very sensitive to
> cut, and further produces the great effect on the normalized histogram.
>
> Best
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Revision history for this message
joholi (joholiphy) said : 		#4

Thanks Olivier Mattelaer, that solved my question.