Top 5 Plays of the Super Bowl? Market says No.

Someone pointed out to me that Protrade has posted their top 5 impact plays of the Super Bowl. But note that they derive the probability impact of each play from an unspecified algorithm. This algorithm estimated that the Bears had a 50% chance of winning at the start of the game, which is far from the probability implied by the Las Vegas odds. These Las Vegas odds have proven over an over again to be as accurate as any other measure. My analysis uses a market-based measure of the probability impact of each play, which builds in how a play affected the market’s expectation of future plays. For example, Protrade’s analysis estimates that Hester’s TD return (their 2nd highest impact play) gave the Bears a 70% chance of winning (20% increase), whereas my analysis estimates that the Bears had a 42.75% chance of winning (10.25% increase, just out of my top 5) following the TD. I think the market correctly anticipated that the Colts would not continue to kick to Hester, so the TD amounted to just 7 points and not much of an indication how the rest of the game would play out.

I suspect that someone who tried trading in a real money market using the Protrade probability estimates would have lost a heck of a lot of money during the Super Bowl (though admittedly this is a small sample of just one game).