[scatters]

There is no benefit to a gene in benefitting the species per se. A trait that increases the size of the gene pool while decreasing its own absolute prevalence will die out.

Group selection is a possible mechanism in social animals, but it requires the group to be strongly interrelated relative to the rest of the species, and it relies on the group being able to expand at the expense of the rest of the species.

Infanticide makes perfect sense; offspring are only related 50% to the parent, so if resources are less than 2x as effective in the bodies of offspring compared to once they have been digested by the parent, of course they should be consumed. The genes of the offspring won't agree, of course; for them the calculation is reversed.

[unishark]

I didn't say the trait would decrease its own overall prevalence. The trait wasn't simply to be depressed if get the genes, it was a predisposition to being depressed if circumstances not good for reproduction. If resources are limited, it would simply ensure that more successful members of a species are the ones to reproduce and get them. Same as in your infanticide calculation.