Comment 130 by ccw95005 :

Good. You have started to define a model of group selection from which a computer simulation could be written to demonstrate the phenomenon, or a mathematical model could be created to prove it.

But you haven't specified everything:-

1) How does each allele in each individual change the probability of survival of each member of the group? For instance, you could state that each group consists of 15 female-male pairs. They each have 4 children, 2 male, 2 female. The survival of each child is 50%, plus 1/2*1/280 for each "cooperative" allele in the 28 adults, at 10 loci, of the group members who are not its parents. I.e. if the whole group is hyper-cooperative the probability of survival of each child is 75%.

Secondly, and more importantly, how do you model the difference between selection of groups, and selection of kin? If the children stayed in the same group and randomly mated with other members of the group, co-operative alleles would soon meet copies of themselves in kin. In a model, one could pool the children from all groups, and then randomly assign them to new groups every generation. As far as I know that is the similar to the "haystack" model, but the wiki description implies that the random mixing only happens every n generations where n is not 1. You could vary the mixing so that each child had an enhanced probability of staying in its own group, or that females stayed, and males randomly found another group, etc.

Write the code, and see what happens!