auto decay width for off-shell case

Hi,

The auto-width mode is only about computing the width and the result of the computation (i.e. the value of the width) does not depend of the process under-consideration.

Now they are limitation to the computation of the width and each time you use the auto-width we remind you of those limitations
- Narrow-width approximation is assumed (so this assume that no particle are degenerate (or close to be degenerate)
- We only compute partial width contribution which are tree-level (so scalar decay to photon/gluon via loop-induced diagram are not included and this can lead to unphysical result -- this is in particular true for the Higgs--)

> Because if I set the width auto, I always got this warning, and I am not sure what madgraph will do with the off-shell region...

Yes you will always have this warning.

> For example if I generate p p > l1 l2 with a s-channel BSM particle mediated. I think this syntax both allow on-shell and off-shell BSM particle. And should I set the width auto just like if I do p p > newparticle, newparticle > l1 l2?

They are no difference of treatment between setting a value for the width or setting the width to auto.
In the auto mode the value is first computed and then both computation will perform in the exact same way.

Cheers,

Olivier

> On 20 Aug 2021, at 19:55, Luanqibazaozc <email address hidden> wrote:
>
> New question #698447 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/698447
>
> Hi, I am a new user of MadGraph. I have read the tutorial https://cp3.irmp.ucl.ac.be/projects/madgraph/attachment/wiki/Shanghai2015/Tutorial_shangai_basic_mg5.pdf
>
> But I am confused about the warning "Please note that the automatic computation of the width is only valid in narrow-width approximation and at tree-level."
>
> For example if I generate p p > l1 l2 with a s-channel BSM particle mediated. I think this syntax both allow on-shell and off-shell BSM particle. And should I set the width auto just like if I do p p > newparticle, newparticle > l1 l2?
>
> Because if I set the width auto, I always got this warning, and I am not sure what madgraph will do with the off-shell region...
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Thank you Olivier!

Do you mean that whatever value I set to the width, the result of the computation of p p > l1 l2 is not influenced?

But is there any reason that changing the width influence the cross-section? I have tried the example in the tutorial for the SM process p p > e+ e- and set different values for the WZ and I do get different values for the cross-section.

> Do you mean that whatever value I set to the width, the result of the
> computation of p p > l1 l2 is not influenced?

No not at all.
I mean that the computation of the width (when set to auto) does not depend if you ask
p p > l1 l2 or p p > t t~ or H > b b~

You can have ask to run the auto-width code without any process defined.

> But is there any reason that changing the width influence the cross-
> section? I have tried the example in the tutorial for the SM process p p
>> e+ e- and set different values for the WZ and I do get different
> values for the cross-section.

In presence of resonances, the width control how large is the cross-section.
So they have huge impact on the cross-section.

Cheers,

Olivier

> On 20 Aug 2021, at 21:01, Luanqibazaozc <email address hidden> wrote:
>
> Question #698447 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/698447
>
> Status: Answered => Open
>
> Luanqibazaozc is still having a problem:
> Thank you Olivier!
>
> Do you mean that whatever value I set to the width, the result of the
> computation of p p > l1 l2 is not influenced?
>
> But is there any reason that changing the width influence the cross-
> section? I have tried the example in the tutorial for the SM process p p
>> e+ e- and set different values for the WZ and I do get different
> values for the cross-section.
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Thank you,

I got you. So you are saying that set width auto is just a parallel operation no matter what process I am calculating?

My initial question is actually whether I should set the width auto when I calculate the process like p p > newparticle > l1 l2 (I think this syntax will both generate on/off shell newparticles)
and the warning makes me a little bit confused...

Because I was thinking when MG calculating this process, when the newparticle is on -shell, the cross-section should proportional to 1/width in the NWA and we need to use auto width to get the correct value. (Is this understanding correct?)

And when the newparticle is off-shell, MG will still use the same width and just do the whole integral?

> I got you. So you are saying that set width auto is just a parallel
> operation no matter what process I am calculating?

correct. I would have call it a pre-processing operation.

> My initial question is actually whether I should set the width auto when I calculate the process like p p > newparticle > l1 l2 (I think this syntax will both generate on/off shell newparticles)
> and the warning makes me a little bit confused...

> And when the newparticle is off-shell, MG will still use the same width and just do the whole integral?

MG5aMC will always do the full integral both for the onshell and offshell regime (otherwise cuts will not be possible).
When NWA is valid, you will get your 1/width dependence from the integration obviously.

Cheers,

Olivier

> On 20 Aug 2021, at 21:35, Luanqibazaozc <email address hidden> wrote:
>
> Question #698447 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/698447
>
> Status: Answered => Open
>
> Luanqibazaozc is still having a problem:
> Thank you,
>
> I got you. So you are saying that set width auto is just a parallel
> operation no matter what process I am calculating?
>
>
> My initial question is actually whether I should set the width auto when I calculate the process like p p > newparticle > l1 l2 (I think this syntax will both generate on/off shell newparticles)
> and the warning makes me a little bit confused...
>
>
> Because I was thinking when MG calculating this process, when the newparticle is on -shell, the cross-section should proportional to 1/width in the NWA and we need to use auto width to get the correct value. (Is this understanding correct?)
>
>
> And when the newparticle is off-shell, MG will still use the same width and just do the whole integral?
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Thank you very much!

I have just one follow up question.

Since MG is doing the full integral, If I am searching for a large width (10%~20%m) newparticle, is the cross-section reliable if I calculate p p > newparticle > l1 l2 (no SM process here, so it should be same as p p > l1 l2)?

And on the other hand, the cross-section is not good if I calculate p p > newparticle, newparticle > l1 l2 because the width is too large to say newparticle is on-shell?

Thank you Olivier,

I am typing "> newparticle >" just to emphasize the mediator not to select particular diagrams. I understand that the inteference with other possible diagrams could be important.

Thank you again!

Thanks Olivier Mattelaer, that solved my question.