Doubting Didymus

Something I've often wondered about is when people say this or that would take 500 000 generations to develop. My question is: is this estimate based on a dircect line single lineage?

i.e. is the author assuming for example, that if there is a 1% chance of a positive mutation, it will take 100 generations to see it?

If this is the case, then I can see the problem straight away. What if the generation is bigger than one man? If a breeding population contains 100 fertile individuals, and each produces ten childern, then that's 1000 chances at hitting the positive mutation in a single generation.

Actually, now that I think about it, seeing as the lucky few who have positive mutations are most likely to survive childbirth, we can take the set of 'eligible mutators' to be every child produced, irrespective of whether they survive. For each human female (in prehitoric times) this number is closer to 30 or so. So in a population of 1000, a single generation might have around 30,000 chances of hitting a positive mutation, even if the total number of survivors is just at the replacement level (1000). When you consider the high likelihood that the good mutants are the ones who will survive in a generation and breed with each other (thus spreading the mutaion throughout the entire population), we are no longer looking at a single lineage with a mutation every blue moon, but using this kind of model instead our probabilities are vastly increased.

I am under the impression that this (in a less simplified form, of course) is exactly what most mathematical mutation representations are doing. Is it just that judith hasn't done her homework?