Comment 4 for bug 1960966

So the issue is that this optimization of MG5aMC assumes that the name of the lorentz structure is based on the spin type of the particles.
You do you have one for which this is not the case (FF1).
if you change that name to FFVFF1 then this fix the problem.

This is indeed an issue of that optimizer and I might try to use a more secure version to get the spin of each interaction if this occurs too often. But for the moment I will just close the bug report.

Thanks a lot for the model, this is very usefull.

Cheers,

Olivier