Update Likelihood for new Ambiguity type from !35293

There is a rather long comment on the code.

The issue is that now we have an additonal type due to !35293 (merged) which we want to keep for electrons

The other part of the issue is that the "cut" is on config txt in GroupArea, and the criterion is 31 ... i.e a hard coded magic number not in "code"

Obviously a better solution can come with the new tunes, when we will update configs. Then I hope my long comment in the code will get removed, as the tuning will progress after the freeze and validation.

For now doing a best effort patch and adding a comment.

Mentioning @turra , @psommer , @okiverny @dboerner , @ludovica