Intrade Fallible?

I have been simulating Intrade prices on Election Day and then assuming Intrade markets are perfect predictors of the actual election. In fact, Intrade does have a perfect record in US elections if the price is outside of the range of [.3,.7]. Inside that range, I'm going to assume there is a chance Intrade is wrong. In fact, I'm going to assume the price on Election Day (when between .3 and .7) is the probability of winning. Further, I'll assume Intrade's fallibility is not correlated from state to state, but simply an independent source of error. That means Obama now has a 96% chance of winning. His lead is dominant, time is short, and McCain probably doesn't have a game changer up his sleeve.