If we really want to go for realistic gender representation, then we should consider two simple rules:
- Survival of the fittest
- Might makes right
What I mean with that is, that the more primitive the culture is, the less it will care about what´s “right” and more about what´s efficient in the means of survival. If you are a nice stone age girl with a biceps the size of my head, I wouldn´t mind if you´d come along for a nice fight with the other tribe.
So how about giving stats requirements for the jobs of the community. E.g. accuracy for hunting or something like that. And then if I need 10 hunters, the AI gives me the 10 people with the highest ACC in the tribe.
The other thing is, that, if we really want to make it accurate, then we have to consider the biological realities of the two sexes. Men in general are stronger. To implement this I´d imagine randomized stats with a different starting point for the two sexes. Let´s say the strength of men are 50+/-25 and the women´s STR is 40+/-25. This way the general superiority of men in STR is represented, but does not exclude the women from STR dependent jobs. Lets say I have 10 men with a STR of 35, 36, 40, 46, 50, 58, 64, 66, 68, 73 (I used an online random number generator ) and 10 women with 10 STR lower stats. If I need 10 people to chop wood (requires STR), that would mean 6 men and 4 women.
Obviously, we would have to consider a lot of other factors to (like one is sick, the other injured, the third is pregnant, etc.), but how about that as a general idea?