Calculation of cross-section using event weights

Hi,

I guess that you have a normalisation issue.

When you split the sample in three sub-sample, you have to adjust the cross-section (equivalently the weight of each event) by the ratio of accepted number of events in that sample.

A more clear explanation is to first change the normalisation of the weight such that they are normalised to the sum.
(i.e. the sum of the weight is equal to the cross-section instead as the average)
In that case, you can then do what you did.

Cheers,

Olivier

PS: You can change the normalisation of the weight by setting in the run_card.dat event_norm to sum
> On 31 May 2017, at 17:14, Ramkrishna <email address hidden> wrote:
>
> New question #637904 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/637904
>
> Dear experts,
>
> As I know using the function "XWGTUP", we can extract the event weight. And the average of weights gives the cross-section. If I am right then I have an issue else please let me know if there is another way of which using the information given in <event> ... </event> patch, how I can extract the cross-section.
>
> I generated the polarised sample using the following method:
>
> 1. generated inclusive p p > W+ W- JJ sample.
> 2. Separated this LHE file in three different LHE file (LL, LT, and TT) based on spin information.
> 3. Finally, I decayed them using the "DECAY Package" (found in older version of Madgraph).
>
> After doing all this, I calculated the average of event weights for this and my results are:
>
> Cross-Section for WLWL = 0.6913 pb
> Cross-Section for WLWT = 0.4197 pb
> Cross-Section for WTWT = 0.5178 pb
>
> But, as we know that the cross-section should follow order:
>
> WLWL < WLWT < WTWT
>
> So, can I safely seperate the LHE file as I described above? If you observe any mistake or you need more information then please let me know.
>
> Thanks & regards,
> Ram
>
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.