Systematics with alternative dyn scales and FxFx merging

Asked by Hannes on 2020-11-05

Hi,

we noticed that when setting ickkw=3 no weights for alternative dynamic scales are calculated by the python systematics module, even though they are set with the --dyn argument.
Is this a bug or due to limitations in combination with FxFx merging? We didn't see any warning or explanation.
This has been tested with v2.7.3.

Cheers,
Hannes

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
Rikkert Frederix Edit question
Solved by:
Hannes
Solved:
2020-11-09
Last query:
2020-11-09
Last reply:
2020-11-09

This question was reopened

Hi,

FxFx merging is associated to his own scale choice (as it is done for MLM merging).
in order to have smooth transition with parton-shower.

So indeed in systematics.py we do have the following lines:

        # For FxFx only mode -1 makes sense
        if matching_mode == 3:
            self.dyn = [-1]

If you comment thoose lines then you will have the output that you expect.
Now I will assign this to Rik (one author of FxFx) to double check with him if he thinks that we should veto or allow this possibility.

Cheers,

Olivier

> On 6 Nov 2020, at 00:05, Hannes <email address hidden> wrote:
>
> New question #693848 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/693848
>
> Hi,
>
> we noticed that when setting ickkw=3 no weights for alternative dynamic scales are calculated by the python systematics module, even though they are set with the --dyn argument.
> Is this a bug or due to limitations in combination with FxFx merging? We didn't see any warning or explanation.
> This has been tested with v2.7.3.
>
> Cheers,
> Hannes
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Rikkert Frederix (frederix) said : #2

Hello Hannes and Olivier,

Changing the default functional form for the renormalisation and factorisation scales will (formally) reduce the accuracy of the predictions in FxFx and therefore not advised.

Best,
Rikkert

Hannes (hannes3) said : #3

Hi Rikkert and Olivier,

thanks a lot for the explanation.

Cheers,
Hannes