Sexual selection perhaps (the ugly ones don't breed). Lets take an
example. Assume a population of say 5 billion people. This population
is not-isolated, and there are no small enclaves (i.e. people can and do
frequently marry people who are distant from them). The Average female
is able to bear up to 10 children, with three being about average. Now
assume this one female has a super-good-awesome-mutation, that allows her
on average to have up to 200 children with the average having 100.
The descendants with super-good-awesome-mutation will always only be a
small fraction of the population (but they never get isolated and are
always capable of interbreeding with average people). They will never
overtake and begin to replace the average people. This is similar to a
math problem I had in the eleventh grade: person A is putting $100 into
an IRA account once a month (say 8% interest) thirty years later person B
is putting $1000 into an IRA account once a month. When will person B's
total accumulated IRA account bypass Person A's (they keep putting in
indefinitely)? Never. The same thing happens in the population
scenario, the better genes are in only a small fraction of the total
population. The same reason that the six finger gene is extremely rare
in the current population.
Uh I think this is wrong although I am math challenged.
The gene would in fact spread in the population rather quickly under the circumstances you describe. Such experiments do happen in nature. Resistance to pesticides is a good example. A gene that confers resistance becomes essentially universally present within a few generations. The problem with large populations is not as far as I know that variation adaptation does not occur but that the wholesale changes in the genotype necessary produce speciation (in particular reproductive isolation) rarely occur in large single populations.
