We've had some attempts to hash it out (for example, comparing round-robin games of chicken to evolutionary equilibria, with niches for a whole spectrum of agents that win against more-cooperative agents but lose against more-defecting agents). Or talking about Schelling points. But I don't think there's been anything really definitive - you should share :)
If it's worth saying, but not worth its own post (even in Discussion), then it goes here.