Weight ordering for systematics

Hi Zach,

> could we assume that the 15th weight will always have the same physics meaning?

The order is not define in any convention, so in principle it can change at any time.
What is more likely to happen is that, one day, someone add some weight before those already define (which is perfectly valid).
Of course, if I would encourage to add those after the already define one.

So to my point of view, you should at least check that the definition of the weight (present in the header) is the expected one.
After that you can use those as you want.

Cheers,

Olivier

On 12 Jun 2015, at 12:01, Zachary Marshall <email address hidden> wrote:

> New question #268071 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/268071
>
> For the same job configuration, is there any way that MadGraph5 / aMC@NLO's systematic variation event weights could change order? For example, assuming we have exactly the same setup, same process, same specification for the process, run, and param cards, etc, could we assume that the 15th weight will always have the same physics meaning?
>
>
> Thanks,
> Zach
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Thanks Olivier. I understand that if we change anything significant (MadGraph5 version, process, Pythia8 version, etc) then things could potentially change. But do you really mean that if we change random number seeds - or even randomly from one job to the next with the exact same configuration - the weights can appear in a different order?

Thanks,
Zach

That should be ok.

Cheers,

Olivier
On 13 Jun 2015, at 18:16, Zachary Marshall <email address hidden> wrote:

> Question #268071 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/268071
>
> Status: Answered => Open
>
> Zachary Marshall is still having a problem:
> Thanks Olivier. I understand that if we change anything significant
> (MadGraph5 version, process, Pythia8 version, etc) then things could
> potentially change. But do you really mean that if we change random
> number seeds - or even randomly from one job to the next with the exact
> same configuration - the weights can appear in a different order?
>
> Thanks,
> Zach
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.