MadGraph5_aMC@NLO

MadGraph event weight

Asked by Guangyi Zhang

Dear expert,

Currently I'm using MadGraph5 to generate ZZ(llnunu) +2 jets EWK LO sample, but I'm confused with the event weight. For example, I used three different seeds to run three parellel jobs, if I looked at the LHE files, the events in the same job have a same event weight, but the event weights in different jobs are slightly different and also very small: weight_job1=0.2449818E-06, weight_job2=0.2477091E-06, weight_job3=0.2456909E-06. If I transfered the output EVNT gen file after parton shower to the DAOD_TRUTH1 format and printed out the event weight, I found the event weights in the same job are different. Taking job1 as an example: weight=0.252366E-06, 0.233027E-06, 0.269852E-06, 0.246637E-06, 0.23398E-06......
How to understand these mc event weights? Are they reasonable? Thanks.

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
Best Olivier Mattelaer (olivier-mattelaer) said : 		#1

Hi,

> but the event weights in different jobs are slightly different and also very small:

The weight correspond to the cross-section of the sample (up to an arbitrary normalization, in madgraph you can either decide that the sum of the weight is equal to the cross-section or the average of the weight. Depending of the MG5aMC version (and of your parton-shower) the default normalization might be one or the other).
Therefore since each of your run returns a different cross-section, you have different weight for each of the three sample.

> If I transferred the output EVNT gen file after parton shower to the DAOD_TRUTH1 format and printed out the event weight,

This I have no idea of what you are doing, You are probably playing with the output of a parton shower and transfer it to a non standard root format.
Since none of them are handle by MG5aMC@NLO, I can not comment on that.

> I found the event weights in the same job are different

This probably means that you are not propagating the correct weight but one weight corresponding to the one of the scale variation.
(or PDF variation). please check that you pick the correct weight.

Cheers,

Oliver

> On Dec 9, 2016, at 18:08, Guangyi Zhang <email address hidden> wrote:
>
> New question #406593 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/406593
>
> Dear expert,
>
> Currently I'm using MadGraph5 to generate ZZ(llnunu) +2 jets EWK LO sample, but I'm confused with the event weight. For example, I used three different seeds to run three parellel jobs, if I looked at the LHE files, the events in the same job have a same event weight, but the event weights in different jobs are slightly different and also very small: weight_job1=0.2449818E-06, weight_job2=0.2477091E-06, weight_job3=0.2456909E-06. If I transfered the output EVNT gen file after parton shower to the DAOD_TRUTH1 format and printed out the event weight, I found the event weights in the same job are different. Taking job1 as an example: weight=0.252366E-06, 0.233027E-06, 0.269852E-06, 0.246637E-06, 0.23398E-06......
> How to understand these mc event weights? Are they reasonable? Thanks.
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Revision history for this message
Guangyi Zhang (guzhang) said : 		#2

Thanks Olivier Mattelaer, that solved my question.