I've really enjoyed reading your work. I'm a physicist, so the very idea that we can generate probability distribution functions for human actions is just mind blowing. I'm actually starting to learn some game theory because of it. Anyway, I have one question - do you think meta-predictions could counter attempts to game your system? For example, we could start a prediction market that predicts a consensus of voters will vote against the truth in one or more markets. The reason I ask is that Nash equilibria are supposed to make law enforcement unnecessary (For example, in the stop light game it's unnecessary to have a law against running the stop light because players have enough incentive not to). In the same way, doesn't the protocol act like law enforcement, and if so could such "second-order" predictions offload some of the work of ensuring players remain honest?