I think you are looking at it incorrectly. I'm not applying anything I'm saying to Screw here at all, but you cannot use recruitment policy to say someone is a bad commander. Someone can recruit and still be a great commander but choose to minimise risk instead.
True, and as I've said before, it's impossible to get a 100% accurate view of a leader from the outside. By definition, I will hear more from the less happy Screw members than from the very happy ones. The same self-selection bias applies within a tribe, where the people who stuck around are more likely to be positive about their leader than the ones who didn't.
I'll happily admit I'm trying to form an assessment based on incomplete data...and so is everyone else That doesn't make it wrong or pointless, simply hard/impossible to prove.
EDIT: I'm using Screw as example because this is the only world I played in the past 5 years. I'll happily continue this with tribes from W16 or W45 as an example, I just doubt most people would be able to add anything to that.