The Iowa futures market now has Obama with an 87% chance of winning in November. The bettors at Intrade give Obama the same odds (even though there is evidence that for awhile someone was trying to manipulate the Intrade market to give McCain better odds; the manipulation failed, and some traders made a nice piece of change taking advantage of the attempted manipulation.)
On a more local level, looking at the Minnesota Senate race, the polls show its a close race between Rep. Norm Coleman and De, Al Franken. But futures traders on the Iowa market seem to think the polls are too timid. Traders give Franken nearly 67% odds of winning...
When the election is over, there should be some interesting academic studies into the different messages of the polls and futures. I'm no expert, and I only check-in occasionally, but it seems that the futures markets have been saying it hasn't been a close race for some time. The polls keep reading it as a close race and there has been a lot of ink spilled on whether there will be a Bradely effect or not.
My guess is that the presidential prediction markets are better indicators of what will actually happen in the voting booth.