As you can see, the best indicator of electoral success comes in the lead up to the election, somewhere between three and four weeks out. That is when things are settled, there is a “time factor” included in the pricing and more than anything, people are honest about their intention yet open minded enough to actually put their money where their mouth is.
Crazy people got on Kerry in the last few days until he was favourite. Only in the last few minutes have people actually realised Bush is going to win and he shot to favouritism.
Not that I am an avid gambler, but I am more interested in studying the proposition that the flow of money in risk managment is often a far better predictor of outcomes than an opinion poll or other such measurement which is open to bias, manipulation and media spin. One of the most recent academic papers has predicted a Bush victory (albeit narrowly) based purely on the flow of money in betting markets.