MadGraph5_aMC@NLO

Reweighting to default or custom dynamic scale

Asked by Kenneth Long

Hi experts,

I've implemented a custom dynamic scale that I'd like to compare to the MadGraph default. I'd love to be able to do this in a single sample with event weights. However, when I use dynamical_scale=0 as the central scale and then reweight with

systematics run_01 --dyn=-1,1,2,3,4

It appears that the -1 option is interpreted as the central scale choice, regardless of whether this was really 0. If I instead produce the sample with dynamical_scale = -1 and then try to reweight with --dyn=0,... the code crashes with a key error from the systematics module. However, if i fix the initial dictionary lookup problem I still see further errors. From glancing over to code, it seems like this isn't really supported.

Am I missing a more straightforward way to do this? If not, is there any plan to have this supported? The process I'm looking at is p p > w+ l+ l- j j QED=6 QCD=0, but I'm suspect this is process independent.

Thanks!

Kenneth

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Solved by:
Kenneth Long
Solved:
Last query:
Last reply:
Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#1

Hi Kenneth,

For the systematics program that i have wrote “-1” has a different meaning of the one in the run_card.dat
For this script “-1” means central scale (the one associated to the event) wathever the formula is to get those scale.

The module systematics.py is actually not able to evaluate the “default” madgraph scale at LO for a given event.
Actually, I do not think that the LHE file has enough information in order to compute that scale a posteriori.

The only solution that I see (I have not checked that it works actually)
would be to update systematics.py around line 752
in order to add your new scale in it.

Recovering the definition from the fortran code would be quite challenging if we do not re-factorize the code.
But I like the concept and keep it for later.

Cheers,

Olivier

> On Nov 2, 2017, at 11:59, Kenneth Long <email address hidden> wrote:
>
> New question #660250 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/660250
>
> Hi experts,
>
> I've implemented a custom dynamic scale that I'd like to compare to the MadGraph default. I'd love to be able to do this in a single sample with event weights. However, when I use dynamical_scale=0 as the central scale and then reweight with
>
> systematics run_01 --dyn=-1,1,2,3,4
>
> It appears that the -1 option is interpreted as the central scale choice, regardless of whether this was really 0. If I instead produce the sample with dynamical_scale = -1 and then try to reweight with --dyn=0,... the code crashes with a key error from the systematics module. However, if i fix the initial dictionary lookup problem I still see further errors. From glancing over to code, it seems like this isn't really supported.
>
> Am I missing a more straightforward way to do this? If not, is there any plan to have this supported? The process I'm looking at is p p > w+ l+ l- j j QED=6 QCD=0, but I'm suspect this is process independent.
>
> Thanks!
>
> Kenneth
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Revision history for this message
Kenneth Long (kdlong-e) said : 		#2

Hi Olivier,

Thanks a lot. Actually I'm pretty happy with this setup, if I'd have realized this earlier it would have saved me the scary experience of having to write some fortran.

Best,

Kenneth