Cuts in madspin

Hi,

One issue is that g g > h is (depending of the model) not a tree-level computation.
so you have plenty of way to handle such computation and the answer to your question depend of the method that you used

1) use the HEFT model, which add an effective operator such that g g > h is now tree-level computation.
This is a computation that we can even performs at NLO (in QCD) accuracy even if you will loose some mass correction (since this is only valid in the infinite top-mass limit)

2) doing the one-loop computation, but then we can only do a LO computation in that case.
(going to NLO would require two-loop computation which are not available within MG5aMc)

-Depending of the choice that you do above (LO in heft, NLO in heft, LO in the SM (via loop-induced process)) the Madspin option available for the Higgs decay are not the same.

They are only one case where can be implemented within MadSpin. It is if you use the
"spinmode=none" case. (the only mode available if you are within loop-induced).
The spinmode is a very "basic" mode --which is ok in the case of the Higgs.
Where decays are generated in the 1>N and the boosted and merged with the original sample.

In that case any cut can be implemented for the generation of 1>N process. However since those events are generated in the Higgs rest-frame. Which means that you can easily implement a pt cut if your generation is at LO (either in the HEFT or for the loop-induced SM) but this is not possible if you are at NLO due to the presence of one jet for the real contribution.

Hopes it helps,

Olivier

> On 12 Jan 2021, at 15:45, Shubhani Jain <email address hidden> wrote:
>
> New question #694948 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/694948
>
> Hello
>
> We are currently using madspin for producing g g > H at nlo . Is it possible to apply p_t cuts in madspin card on decay product chain in Madgraph, as the cuts defined in run_card don’t seem to be implemented on decay products?
>
>
> Regards
> Shubhani Jain
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Hi Olivier

Thanks for replying back.

We are using 2HDMII_Bol model for this process. Would it be possible to put generation level or cuts for this model in mad spin?

Regards
Shubhani

The default mode of MadSpin use the Branching ratio, in that case you can not apply any cut by construction. If you move to the madspin=none case, in that case you can apply cuts but only in the rest-frame of the Higgs. So since you are doing p p > h and p p > h j (due to NLO) you will not be able to have generation cut at madspin level.

Cheers,

Olivier

> On 12 Jan 2021, at 18:05, Shubhani Jain <email address hidden> wrote:
>
> Question #694948 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/694948
>
> Status: Answered => Open
>
> Shubhani Jain is still having a problem:
> Hi Olivier
>
> Thanks for replying back.
>
> We are using 2HDMII_Bol model for this process. Would it be possible to
> put generation level or cuts for this model in mad spin?
>
> Regards
> Shubhani
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Thanks Oliver for the useful information,

I am a collaborator of Shubhani's, for context we are generating 2HDM events with a H > hh > bbb~b~ decay, and ultimately we are trying to generate a sample with high p_T jets for machine learning. Another collaborator has recommended we should actually cut the Higgs rather than directly the b partons, but any further information on this would be very much appreciated.

Best,
Billy

Having cut on the Higgs (so before madspin) are possible no issue on that.
Depending of the cuts, you might just need to specify those in the run_card (if you want predefined cut)
or need to implement yourself in the code.

Cheers,

Olivier

> On 13 Jan 2021, at 15:25, Billy Ford <email address hidden> wrote:
>
> Question #694948 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/694948
>
> Billy Ford posted a new comment:
> Thanks Oliver for the useful information,
>
> I am a collaborator of Shubhani's, for context we are generating 2HDM
> events with a H > hh > bbb~b~ decay, and ultimately we are trying to
> generate a sample with high p_T jets for machine learning. Another
> collaborator has recommended we should actually cut the Higgs rather
> than directly the b partons, but any further information on this would
> be very much appreciated.
>
> Best,
> Billy
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Just as a sanity check, we are generating g g > h2 in madgraph, then decaying to the two h1's and subsequently each h1 to bb~ pairs in MadSpin. I suspect therefore that applying the h1 pt cut in the run card would not work, as it is not generated until MadSpin is called?

So would we apply them in PROCESS/SubProcesses/cuts.f, or is there a different cut file for MadSpin objects?

Cheers,
Billy

you are correct.
If the decay of h2 is done by madspin then you face all the limitations written above.

The likely solution for you case is to use the syntax
generate g g > h2 > h1 h1 [QCD]
use the run_card to apply a pt cut on h1 (specified via the pdg of h1)
and then use madspin to decay h1

Cheers,

Olivier

> On 14 Jan 2021, at 11:50, Billy Ford <email address hidden> wrote:
>
> Question #694948 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/694948
>
> Billy Ford posted a new comment:
> Just as a sanity check, we are generating g g > h2 in madgraph, then
> decaying to the two h1's and subsequently each h1 to bb~ pairs in
> MadSpin. I suspect therefore that applying the h1 pt cut in the run card
> would not work, as it is not generated until MadSpin is called?
>
> So would we apply them in PROCESS/SubProcesses/cuts.f, or is there a
> different cut file for MadSpin objects?
>
> Cheers,
> Billy
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Oh excellent, thanks very much! I was under the impression that kind of process syntax would not work as decay chains aren't possible at NLO, but seems to have done the trick!

The percentages talk to uncooked variety of MC activities left after the cutflow. As an instance, if one generates 10K occasions, and after the cutflow 1K raw events are left, that percentage could be 10%. We use this information to estimate required statistic tics we want to have in the long run as this project https://thebestcoffeebeans.net.