Have been through that a few times ITT - I don't model chance of winning, I predict market error.
Best analogy would be back like 5-10 years ago in the NFL, you could simply blindly bet on home underdogs and make a profit, because the market would systematically price them too long. But if you asked someone who bet on home underdogs how he thought a particular home dog would be priced before the lines came out, he wouldn't necessarily know. He'd just know whatever number the market ended coming up with would be too long.
Simple betting strategies like that don't work in NFL anymore because the market is too sharp, but the market in MMA is not sharp, to put it mildly. Last card Tuck opened at 2.25 and closed at like 1.30, that right there is not a sharp market. If the MMA market does adjust then a) I am screwed and b) I won't even know until I notice profit start to trail off or can do some analysis showing that the type of fighters I bet are being priced shorter, which is why I'm pretty circumspect about exactly what I am betting on.
Actually modelling MMA would be a tough undertaking and it might be impossible to judge fighter skill better than humans can do it. Even if it could be done it would be so difficult to do that the amount of work wouldn't be worth it given the betting limits. It would also be useless on relatively new fighters there isn't enough data on.
Interestingly the majority of the lines I like tend to shorten from opening, which suggests that the people setting opening lines are less sharp than the public (which is maybe not surprising in a sport that is hard to model). There is (on average) still value left when the lines close, in fact lines shortening is in itself a predictor of value in the closing lines. I've been wondering if this is an example of
anchoring where people aren't really sure where a lot of the lines should be and revise their guesses based on opening lines, but that's speculation.
Last edited by ChrisV; 11-13-2012 at 01:01 AM.