How on-shell Z events are generated or marked?

1. They are two criteria:
 - First the relative importance of each diagram. Allowing to select one given Feynman diagram.
 - If that diagram has a Z and if that Z is in the onshell window then it will be written.
Now if you want to have more details on the exact algo, you can read this question: https://answers.launchpad.net/mg5amcnlo/+question/819446

2. Maybe I do not know enough that process to answer such question. But given that the width divided by the mass is also around 2 % this sounds reasonable.

Cheers,

Olivier

Dear Olivier:

Thanks for your reply. I apologize for not realizing that my colleague
had asked a similar question. After discussing with him, I still have
some confuses:

1. About the sampling
    I read the MadEvent paper about SDE strategy, which mentions a
    g_(i)(Φ) density function. Should the Breit-Wigner distribution be
    considered a specific choice of g_(i)(Φ)? As you specified:

      determine if the events needs to written on disk or not
      (unweighting stage is based on the maximum of the integrand
      –including the multi-channel factor)

    Are you referring to use the acceptance-rejection method to sample
    unweighted events?

2. About the multi-channel factor
    Is this factor introduced solely for the accurate evaluation of
    phase-space integration? Then Z inside mass windows in Z channel is
    just defined as on-shell Z and we can hardly reconstruct same Z with
    only final state leptons?

3. Compared with other on-shell Z contributions from
    1. on-shell Z events + MadSpin
    2. decay chain syntax in MadGraph
    are those Z events more physical or more acceptable?

Best,
Shux Li

Hi,

1. Are you referring to use the acceptance-rejection method to sample
    unweighted events?

Yes

2. About the multi-channel factor
    Is this factor introduced solely for the accurate evaluation of
    phase-space integration?

Can not really comment on the intent of the author of that paper.

3. Then Z inside mass windows in Z channel is
    just defined as on-shell Z and we can hardly reconstruct same Z with
    only final state leptons?

This clearly gives a "simple" way to decide how to write propagator. But nothing prevent you to have different methods.
(potentially better one)

3. Compared with other on-shell Z contributions from
    1. on-shell Z events + MadSpin
    2. decay chain syntax in MadGraph
    are those Z events more physical or more acceptable?

Those are really similar, the only difference is the total cross-section which is better in two compare than the one from one, since MadSpin use branching ration while "2" does phase-space integration.

Cheers,

Olivier

Hi Oliver:

Thank you for patient reply. How e+ e- $ z events are generated?
Compared with directly remove Z events from e+ e- events (with same
iseed, and 1m total events), they differ by 0.05%. But in my case
it's 0.5%.

Best,
Shux Li

This is conceptually the opposite operation,
For any events where we would have written the Z, we change the weight of such event to 0 (which prevent the cross-section to take such event into account and to write such event on disk)

Cheers,

Olivier

So they can differ even with same random seed? I mean I can see exactly same
events (without Z) in e+ e- and e+ e- $ z events, but still some events only
present in either. So are there some unweighting or other procedures to prevent
their exactly match?

Thank you so much for your answer! It helps me a lot!

Thanks Olivier Mattelaer, that solved my question.