I find that these game theoretic models are very limited by their rules. You can pretty much formulate a game to support a hypothesis. I think these type of conclusions can be only be obtained as a data-driven result.