Cross-sections provided by Madgraph+Madspin in dim-8 EFT model

Hello Madgraph Experts,

I am attempting to generate events with a dim-8 EFT (Eboli model http://feynrules.irmp.ucl.ac.be/wiki/AnomalousGaugeCoupling) for the VBS “semileptonic” states ( for example p p > w+ w- j j > l v j j j j) using Madgraph 2.6.7. To speed-up the generation I am producing gridpacks and using Madspin for the final boson decays.

I notice a discrepancy in the resulting cross-sections recorded after running with Madspin with respect to an explicit decay chain calculation. I have used the following syntaxes and received the following values:

a) p p > j j w+ w- QCD=0 T0==1
 cross-section = 0.277 pb
b) p p > j j w+ w-, (w- > l- vl), (w+ > j j) QCD=0 T0=1
 cross-section = 0.0338 pb
c) p p > j j w+ w- QCD=0 T0==1 (followed by madspin in “none” mode for w- > l- vl, w+ > j j)
 cross-section = 0.278 pb
d) p p > j j w+ w- QCD=0 T0==1 (followed by madspin in “madspin” mode for w- > l- vl, w+ > j j)
 cross-section = 0.0618 pb

I am reading the cross-section value from the “Integrated weight (pb)” line of the produced LHE file. It might also be worth noting that for this test a quadratic EFT process was generated (applying the recommended ‘hack’ from https://answers.launchpad.net/mg5amcnlo/+question/255413, where the square-coupling part of the syntax is removed before madspin, as the BSM component should not effect the W/Z decay)

I can see that the recorded value with spinmode=none is the same as the inclusive value so I assume the branching ratio is not applied, but I also notice that the default madspin mode does not match the full madgraph decay chain result.

Regards,
Robert Les