Ok, the kin selection is hypothesis is pretty simple: gay siblings lead to more offspring, especially when competition is strong and resources limited (pretty much always the case for critters like humans)
balancing selection hypothesis is a little more complicated. First, the observation that the biggest determinants of fitness in hominids appear to have been social rather than physical for some time. By the way, this means neither the traits nor the environment fossilize which means that anyone making specific claims about human selection is probably making up fairy stories. Yes, I'm talking about you oh "evolutionary psychologists.") However, if you take a look at the fox domestication experiments, it appears that in mammals there are a series of physiological and hormonal changes that go along with reduced aggression and greater social intelligence. Indeed, it looks like humans have been self-domesticating for some time, with reduced dentition, greater cranial capacity and a few other things. What this suggests is that human selection has been selecting for a greater ability to win and hold allies and empathy and the like (who knew, Diplomacy players are the pinnacle of human evolution!). Now, the hypothesis is that there's a developmental correlation between those traits and sexual orientation through some common developmental pathway, particularly among males. Sure, you can pump out maximum testosterone and be super hetero, but at a cost of being piss poor at crafting the alliances that are needed for greater reproduction. So, hypermasculinity is selected against. On the other hand, if emotional intelligence is part of a complex of traits that includes some of the biological foundations of homosexual sexual orientation, then you'll have an instance where maximum male fitness comes from being "a little bit gay" That comes at a cost of some offspring with reduced fitness, but that's offset by greater fitness from the others. Be too hypermasculine and fail. This is somewhat analogous to the mechanism that keeps sickle cell anemia around at low levels. Heterozygotes are immune to malaria, and so have greater fitness. However, if they become too common in a population, more of your offspring are homozygotes and die quickly. So, there's a frequency dependent selection there to keep the trait around, but not too common.
Anyway, the point is that evolution is really bloody complex, and probably more complex than hominid brains can really understand. Its not some simplistic thing like physics. That in turn means that everyone ought to be really damned careful about making evolutionary claims. Otherwise you end up looking like an idiot like the "sex is for reproduction" religious guys who clearly understand precisely zero about humans. You don't want to be end up a laughingstock like those guys.