MadGraph generates fewer events than the nevents I set

[Ren Hong-Yu]

Hi,

When I generate the process
"u u > w+ w+ d d QCD=4"
with the latest version of MadGraph5, 1.5.12, I found it gave fewer events than the nevents (=10000) I set in run_card.dat.

I did not change anything of the package and just used the sm. I met this problems many times for almost all the versions. It has confused me for a long time.

Thanks for your time!

Best Wishes!

Hong-Yu

[Valentin Hirschi (valentin-hirschi)]

Dear Ren,

This is a known behavior of VBF process with MG5_aMC.

The basic explanation for this is that the topology of the process is a bit peculiar and the automatic multi-channeling integration technique used in MadEvent does not map the integrand very well. As a result of this, the event generation efficiency is too low and doesn't reach the 10k events target with the usual number of points / iterations specified by MadEvent.

There is no easy immediate fix for this (besides automatically running for longer until it reaches the target number of events, but we chose not to do so because if the 10k events are not reached, it usually signals a problem).
Olivier Mattelaer might be able to comment more on this and the possible solution to anticipate in future version (if any). He is not available for the moment but will get back to you when he'll be back.

[Olivier Mattelaer (olivier-mattelaer)]

At least now, this is providing the correct number of events for the process specified above.
Such kind of problem are still expected for VBF but the situation improved quite a lot already.

Cheers,

Olivier