Madanalysis 5 could not read lhco file correctly

Hi Karo,

Could you please provide me the script you have typed in?

Thanks in advance,

Benjamin

Hi Benjamin,

    I input the command:

      ./bin/ma5 -R
       import sample/delphes.lhco
       plot MET 60 0 600 [logY]

Cheers,
Karo

Hi Karo,

The cross section is not present in the LHCO file so that MadAnalysis 5 is using the default value of 0. Therefore, the normalization is set to zero too and it is impossible to normalize the figure to the luminosity. By making use of "set defaultset.xsection = ...", your issue should be fixed.

Cheers,

Benjamin

Hi Benjamin,

   I have set "set defaultset.xsection=...". Other things is correct, eg,
  Generation: 20000 events.
* Cross section imposed by the user: 1.52 pb.
* Normalization to 10 fb-1: 15200 +/- 0 events.
* Ratio (event weight) = 0.76 .

   But the y-axis of histogram is still "events(not normalized)".

Cheers,
Karo

Hi Karo,

Have you added to your script "set main.normalize = lumi"?

Cheers,

Benjamin

I have added the set of lumi. It works well.

Thanks very much
Karo

Hi

   I have a question again. The cross section which I must set is before matched cross section or after matched cross section? The events after matching is not 20000. There are some events not passing matching. However, the LHCO file has 20000 events, but some events is empty.
   I am so confused.

Cheers,
Karo

Hi Karo,

You need to use the cross section after matching, and don't worry about the empty events. This is a delphes' thing. MA5 should tell you the right number of events.

Cheers,

Benjamin

Hi Benjamin,

    In the MA5 report, the first table shows,

      event files number of events cross section
    myfile.dephles.lhco 20000

     So I am worry about the cross section. And I think it maybe is before matching cross section. Otherwise, how it shows 20000?

Thanks,
Karo

Hi Karo,

I think that this is a Delphes feature. Olivier could possibly confirm. Among your 20.000 events, a bunch of them a just zero. And they therefore do not contribute to the histograms. Consequently, the proper xsection to use is the one after matching.

FYI, MadAnalysis is counting 20000 events in total because you have 20000 events in the sample. What you could do to convince yourself is to plot any global observable which is not zero (e.g. the HT variable) and see the size of the zero bin.

Cheers,

Benjamin

Yes indeed this is a feature of Delphes,

The point, I think, is to keep empty events such that it is easy to trace the history of the events.

Cheers,

Olivier
On Apr 26, 2013, at 6:51 PM, Benjamin Fuks <email address hidden> wrote:

> Question #227248 on MadGraph5 changed:
> https://answers.launchpad.net/madgraph5/+question/227248
>
> Status: Open => Answered
>
> Benjamin Fuks proposed the following answer:
> Hi Karo,
>
> I think that this is a Delphes feature. Olivier could possibly confirm.
> Among your 20.000 events, a bunch of them a just zero. And they
> therefore do not contribute to the histograms. Consequently, the proper
> xsection to use is the one after matching.
>
> FYI, MadAnalysis is counting 20000 events in total because you have
> 20000 events in the sample. What you could do to convince yourself is to
> plot any global observable which is not zero (e.g. the HT variable) and
> see the size of the zero bin.
>
> Cheers,
>
> Benjamin
>
> --
> You received this question notification because you are a member of
> MadTeam, which is an answer contact for MadGraph5.

 Hi Olivier,Benjamin,

Thank you very much! I have understand it.

Cheers,
Karo