LC issue with loop-induced process

[Olivier Mattelaer]

 When generating an event with madgraph like this:

generate c~ g > H g c~ QED=1 [QCD]
launch

if I look into the LHE event card, I get events such as the following:

  <event>
    5 0 +4.0285366e-02 1.36540700e+02 7.54677100e-03 1.21806100e-01
          -4 -1 0 0 0 501 +0.0000000000e+00 +0.0000000000e+00 +2.9419141131e+03 2.9419141131e+03 0.0000000000e+00 0.0000e+00 -1.0000e+00
          21 -1 0 0 501 502 -0.0000000000e+00 -0.0000000000e+00 -2.1835375839e+01 2.1835375839e+01 0.0000000000e+00 0.0000e+00 -1.0000e+00
          25 1 1 2 0 0 +4.5770992300e+01 +2.9989942130e+01 +2.0255992823e+02 2.4423330013e+02 1.2500000000e+02 0.0000e+00 0.0000e+00
          21 1 1 2 501 501 -1.3125965953e+01 +2.3701158420e+01 +3.1121389496e+02 3.1239097987e+02 0.0000000000e+00 0.0000e+00 -1.0000e+00
          -4 1 1 2 0 502 -3.2645026347e+01 -5.3691100550e+01 +2.4063049140e+03 2.4071252089e+03 0.0000000000e+00 0.0000e+00 -1.0000e+00
   <mgrwt>
   <rscale> 4 0.13654070E+03</rscale>
   <asrwt>0</asrwt>
   <pdfrwt beam="1"> 1 -4 0.45260186E+00 0.13654070E+03</pdfrwt>
   <pdfrwt beam="2"> 1 21 0.33592909E-02 0.13654070E+03</pdfrwt>
   <totfact> 0.87394861E+01</totfact>
   </mgrwt>
   </event>

where a gluon singlet state has been produced (with color tags 501 501). Is this expected behaviour? It seems when I try to subsequently shower the event with Pythia, it gives me the error 'unphysical color flow'

[Olivier Mattelaer (olivier-mattelaer)]

See the associate question for more details on this issue

[Olivier Mattelaer (olivier-mattelaer)]

The file leshouches.inc is correct and the assignment of the JAMP2 looks correct as well.
The best idea is a screw up in addmothers.f -> assigned to me

[Olivier Mattelaer (olivier-mattelaer)]

Hi Valentin,

Following our discussiong, the issue is not what we discussed.
If you look at diagram #19 (see attachment)
you will see that the following color-flow:
      DATA (ICOLUP(1,I,3,1),I=1, 5)/ 0,503, 0,503, 0/
      DATA (ICOLUP(2,I,3,1),I=1, 5)/501,502, 0,501,502/

is not possible (at minima weird) since the gluon is not connected to the quark with with it interacts. If this should really be consider as LC, then I would need to correct the addmothers function to correct the issue. But sounds an issue on the LC definition here.

Cheers,

Olivier