Assignment of kinematics at FO

Asked by Claudio Severi

Hello everyone,
this is a (yet another) followup about running NLO FO calculations with a LHE output.

I get in the lhe file some phase space points I did not expect.
For instance, when running the process:
  p p > t t~ QCD=2 QED=0 [QCD]
with all settings left to their default values, i got:

 4 0 -7.7034287e-05 0.00000000e+00 0.00000000e+00 0.00000000e+00
       21 -1 0 0 599 599 +0.0000000000e+00 +0.0000000000e+00 +2.2624600942e+03 2.2624600942e+03 0.0000000000e+00 0.0000e+00 9.0000e+00
        2 -1 0 0 599 599 +0.0000000000e+00 +0.0000000000e+00 -2.3142377606e+03 2.3142377606e+03 0.0000000000e+00 0.0000e+00 9.0000e+00
        6 1 0 0 599 599 +3.7398442496e+02 +7.9711327642e+02 +2.0792477093e+03 2.2645328353e+03 1.7200000000e+02 0.0000e+00 9.0000e+00
       -6 1 0 0 599 599 -3.7398442496e+02 -7.9711327642e+02 -2.1310253757e+03 2.3121650195e+03 1.7200000000e+02 0.0000e+00 9.0000e+00
#aMCatNLO 0 0 0 0 0 0.00000000D+00 0.00000000D+00 9 5 0 0.00000000D+00 0.00000000D+00 0.00000000D+00 0.00000000D+00 0.20000000D+01

The process q g > q q should definitely not be happening here, so I assume there have been some problems with the assignment of momenta or pdg codes. For context, this event is part of an eventgroup, with the counter-event being the expected real-emission process (21 2) -> (6 -6 2).

Can you please confirm what is happening?

Thanks a lot,

Question information

English Edit question
MadGraph5_aMC@NLO Edit question
Rikkert Frederix Edit question
Solved by:
Rikkert Frederix
Last query:
Last reply:
Hi Rik, I doubt that this is specific to the lhe output (looks like this is coming directly out of the ipdg block from the fill_analysis. I guess taht the issue is indeed that the pdg code for initial particle is not correctly set when passed to the "outfun" routine. What do you think?
Revision history for this message
Rikkert Frederix (frederix) said :

Hello Claudio,

This is expected. In the FO analysis, the momenta are always the 5-particle ones, with, in the case of the Born/Virtual/Subtraction contributions, the momentum of one of the particles is equal to zero. When writing the fixed-order phase-space points to file, this momentum is no written to safe space. Indeed, that means that the PDG codes might be somewhat inconsistent. Note, however, this is not wrong: the PDG code assignment is not IR safe, and one cannot naively distinguish quarks and gluons based on the PDG codes in the event file.


Revision history for this message
Eleni Vryonidou (evryonidou) said :

Hi Rikkert,

why is the information of which Born configuration gives the PS point not used to make some "consistent" PDG assignment? Is there a particular reason? Even naively by looking at which Suprocesses folder it comes from the code should be able to associate a Born configuration for each event. Is this way of thinking not correct?



Revision history for this message
Best Rikkert Frederix (frederix) said :

Hi Eleni,

For the Born and virtual contributions this would be okay, but not necessarily for the subtraction terms. The latter have, in principle, one additional parton that is either in the direction of the beams, or soft and/or collinear to a massless outgoing parton. Differentiating between those is a pain, and would significantly enhance the size of the event files, since it would mean writing those contributions as separate events all out in detail. Currently, they are simply grouped together, assuming you apply a IR safe analysis.

Having said that, I guess that for most practical purposes one could simply assign the Born configuration to these contributions. But I don't really see the reason for this, and might give a wrong impression to a user about what these PDG codes mean for partons in a fixed order NLO computation.


Revision history for this message
Claudio Severi (claudio-severi) said :

Thanks Rikkert Frederix, that solved my question.