Madspin with NLO

Hi,

I think that the best is to read the MadSpin paper to have all the details: arXiv:1212.3460

In short, MadSpin use only tree-level matrix-element and the decay part is therefore LO accurate.
This is why we need to pass from the NLO syntax to the LO ones.
The precision of the production part should not be modified by the decay part and is therefore valid for both LO and NLO generation.
(and indeed the code works for both LO and NLO computation)

Your process is actually a LO one but a special one: a loop induced process.
We can not apply the MadSpin algorithm on those matrix-element because madloop is not able to handle loop with decay-chain.
However, we have a special mode in MadSpin for the Higgs decay which can be used with loop induced processes.
(as described in the loop-induced paper: arXiv:1507.00020)

I can see three solutions:
1) generate g g > z h > e+ e- mu+ mu- [QCD]
and then play with the lhe file to change the flavor according to the BR that you want to have all the massless quark flavor.

2) give up on the full spin-correlation by decaying the Z independently. (with either madspin in the “Higgs” mode or making the decay via the PS)

3) A third option is to use the old “Decay” package of Fabio Maltoni. (which is not supported anymore) where you can have a partial spin-correlation.
It is certainly better than having none, but since spin-correlation is so strong for such type of diagram. It is far from clear how good/bad the partial result are.
You can find that package shipped in MadGraph4 tarball.
Do not expect that the code will work out of the box (the lhef format pass from version 1 in MG4 to version 3 in MG5) so be prepare to go inside the fortran and to debug the code yourself since we do not support that code anymore.
I do not know if that code has any official paper related to it. But you should have all the details in a follow-up paper (by some other author)
arxiv.org/pdf/hep-ph/0703031v2

Cheers,

Olivier

> On Dec 6, 2016, at 20:23, Lailin Xu <email address hidden> wrote:
>
> New question #406517 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/406517
>
> Hi,
>
> When I was trying to run MadSpin with a NLO process using loop_sm model:
> generate g g > h > z z [QCD]
>
> I got the following error message:
> Command "generate g g > h > z z --no_warning=duplicate" interrupted with error:
> NoDiagramException : No amplitudes generated from process Process: g g > h> z z WEIGHTED=8 @1. Please enter a valid process^[[0m
> Command "add process g g > h > z z pert_QCD --no_warning=duplicate" interrupted with error:
> NoDiagramException : No amplitudes generated from process Process: g g > h> z z g/d/u/s/c/c~/s~/u~/d~ WEIGHTED=12 @1. Please enter a valid proces
> Command "generate_events run_01" interrupted with error:
> InvalidCmd : No processes generated. Please generate a process first
>
>
> By looking at the code: MG5_aMC_v2_3_3/madgraph/interface/madgraph_interface.py
> I see the following:
>
> def split_process_line(procline):
> """Takes a valid process and return
> a tuple (core_process, options). This removes
> - any NLO specifications.
> - any options
> [Used by MadSpin]
> """
>
> # remove the tag "[*]": this tag is used in aMC@LNO ,
> # but it is not a valid syntax for LO
> line=procline
> pos1=line.find("[")
> if pos1>0:
> pos2=line.find("]")
> if pos2 >pos1:
> line=line[:pos1]+line[pos2+1:]
>
>
>
> What I don't understand is, why it's necessary to remove the NLO specifications with "[QCD]" syntax?
>> From the comment above, does it mean that MadSpin only works for LO?
>
> What's the correct way to use MadSpin in my case?
>
> Thanks,
> Lailin
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.