Prediction Market Mayhem

Sounds like someone is screwing with the intrade contracts. InTrade's contracts on Obama are way, way different from ones at the Iowa Electronic Market and another across the pond, both of which show Obama winning at 61% compared to inTrade, which is closer to 52%. The writer also notes that the individual state contracts don't add up to make it such an even money race for Obama. I computed those state contracts two weeks in a row, and what I saw made it seem like McCain needed to get awfully lucky to pull it off - yet the winner-take-all contract was still at 50/50 or so. Kind of weird. Anyway, it sounds like these prediction markets aren't as efficient as we once thought - there appear to be arbitrage opportunities, but no one taking advantage of them.