Under a sufficiently low population density, genes among splintered groups will diverge. Groups with genes that promote expansion and/or adaptation will eventually out-compete groups that don't. Does this process not occur?
>It's hard to find any examples of genetic material of a group being lost when that group was outcompeted by another group.
Isn't the extinction of other human species an example of this?
Yes, and this occurs.
> Groups with genes that promote expansion and/or adaptation will eventually out-compete groups that don't.
For some definition of "eventually". Selection over genes competes to determine outcomes with "selection" over culture and with chance.
> Isn't the extinction of other human species an example of this?
Yes. And that's a disputed hypothesis.
Granting that hypothesis, how many such gene-selection events occur per ten thousand years? Not many.
How much of observed gene distribution is best explained by selection pressure from these rare events rather than by individual selection operating continuously? Very little if any.