Some few events with high event weights

They should all have the same weights indeed.
Now my advise would be to check the lhef file and the hepmc file to know which output file has the issue.
My guess is that this is Delphes issue (and if this is the case then this should be reported to them).

Cheers,

Olivier

Thank you for your reply!

I checked the "unweighted_events.lhe" file and the pythia8 events hepmc file. Since I am not an expert in these file formats here is a small description of what I checked:

lhe: Using pylhe (https://github.com/scikit-hep/pylhe) I checked eventinfo.weight for the events and it seems that every event has weight 2018.92 (the cross section of the process). Using event.weights (so not eventinfo but just the event) I do not get any weights.

hepmc: Here I used pyhepmc (https://github.com/scikit-hep/pyhepmc) and checked event.weights for each event in the file. This gives 2 weights where the first one seems to be the relevant one because here I get the same problem as in the Delphes root tree: most events have weight 1 and some 2018.92 (exactly the same number of events as in the Delphes root tree).

So it seems the problem occurs already before Delphes.

Cheers,
Cedric

Then the issue is within Pythia8,
I'm not an expert in that file format either.
But I checked an the hepmc produce on my side with the following two commands:

[run_01]$ cat tag_1_pythia8_events.hepmc | grep ^E | grep -v "3.4215000000000004e+00 3.4215000000000002e-06" | wc -l
       0
[run_01]$ cat tag_1_pythia8_events.hepmc | grep ^E | wc -l
 1000000

so all 1 million event have the cross-section as information in my case.

I will not be able to help more on this, my suggestion would be to contact pythia8 authors (who will likely ask you to run outside of MG5aMC framework)

Cheers,

Olivier

Ok, thank you very much for your help.

Cheers,

Cedric