Originally Posted by
mathdotcom
Those not familiar with mathematical reasoning won't be able to understand this anyways. There are a bunch of papers written on this; the simplest form is checking to see if various betting rules would turn a profit, and if so, is it statistically significant. Not surprisingly, blindly betting home underdogs and unders has not proven to be a very successful strategy. The more complicated form, yet extremely simplistic for a publishable paper, uses probit/logit models and simply runs the regression: Prob(Home Win) = Market Estimate of Probability of Home Win + whatever other available information you consider to be important (such as a Home Field dummy, the Total, the previous game's result, etc.) + e and checking to see if any of the other available information is statistically significant. Sadly, the above article would yield a better and more thorough paper than any of the academics have come up with. The famous author of Freakonomics, S. Levitt, uses data from an NFL handicapping tournament to try and make inferences about market efficiency - an idea which is of course ridiculous given tournaments lead to strategic situations that cause bettors to bet games they otherwise wouldn't. Ganch, if you want an article published in a (mediocre) journal you should e-mail some of the current economist dorks working on gambling markets: Andrew Weinbach, Rodney Paul, Woodland & Woodland, Raymond "Skip" Sauer, etc.