W masses when Higgs is off-shell

Hi,

When I'm doing some checks with Higgs decay to WW, by using the HiggsCharacterization model, I see an interesting feature when the Higgs is off-shell. In the parameter card I have the Higgs mass set to 125 GeV and let it decay to W+W- and two leptons and two neutrinos in the final state. So always one of the W has to be off-shell and it's about 50/50 for both the W+ or W- to be on/off-shell in an event. However, I found out that when the Higgs is generated off-shell, the W+ is more often off-shell than the W-, so no 50/50 balance anymore. This seems a bit strange to me, because I don't know any reason for why the W- should be more often on-shell for an off-shell Higgs. The result is also visible in the final state particles, because the l- has in this case on average a higher momentum than l+.

The samples where I see this feature are generated with MadGraph 5 (5.2.0.1). When I compare it to samples generated with aMC@NLO, I see that aMC@NLO doesn't have this feature, so the W are 50/50 off-shell for my events, also when the Higgs is off-shell.

So my question is why does MG5 have a "preference" for the W- to be more often on-shell than W+ when the Higgs is off-shell?

Cheers,
Koen