I used the probabilities from Intrade and did Monte Carlo simulations, a million trials at a time, which showed a 49.4% chance of winning, based on the 12 states I used.
Intrade is just what people’s expectations are, so can be wrong. For example, for a long time they had Obama as a favorite to win in Ohio, even by over 60%. That never made any sense to me. If I were a betting man, I would have taken that, because at no time, in my opinion, did McCain ever have less than a 40% chance of winning Ohio.
As for the Intrade, I trust that more than MSM opinion polls. Before Palin, no one could have predicted the swing McCain has taken. To be fair thought, it isn't over yet, and anything can still happen. I think the closer we get to the election, the less likely we are to see a major shift.