Sagarin

raincat · Post by **raincat** » Thu Feb 17, 2005 10:51 am

So much for the magic computer. Poll shows Bobcats 13 spots below the griz. Jeff, look out the window!!

Cat Grad · Post by **Cat Grad** » Thu Feb 17, 2005 11:03 am

Yes; just read someplace online in the last couple of weeks that were somebody to bet against the poll, they'd be considerably ahead. I'm not sure, but I think it was in Sports Illustrated.

Post by **rtb** » Thu Feb 17, 2005 11:41 am

Cat Grad wrote:Yes; just read someplace online in the last couple of weeks that were somebody to bet against the poll, they'd be considerably ahead. I'm not sure, but I think it was in Sports Illustrated.

Or was it right here on this site? A member posted this in another thread last week.

El Profesor wrote:I thought some of you might be interested in the following information. I have thoroughly researched Sagarin's ability to predict winners of games in NCAA football. I am currently running another study about his ability to predict basketball winners -- early indications are that his winning percentages are similar to those for football. Harmon is another service that publishes power ratings, similar to Sagarin's. The most interesting finding from the following passage is that a wagerer in Las Vegas could have made money by betting AGAINST Sagarin's picks!!! (Remember, the following information pertains to football. Also, the study spans the years 1991-1998, so the results may be different in more recent years, though I doubt they are.)

"Harmon’s ability to predict the outright winners is measured by a 74.39% success rate overall, while Sagarin was 75.55% successful in the later-year games (though only 65.50% successful overall and a very surprising 54.56% for games in 1992-1994). Favored teams tended to win their games 76.71% of the time, suggesting that even Harmon’s success rate is nothing extraordinary, and is actually 2.32% worse than Las Vegas bookmakers’ rates of predicting winners of college football games."

"Against the spread, Harmon is only able to correctly predict which teams will cover the spread half of the time, evidence which is consistent with sports betting markets being efficient. Harmon’s 50.00% overall winning percentage is nearly identical to the 49.18% and 50.17% from the respective naive strategies of always picking the favorites and always picking the home teams. Sagarin’s overall and early-year winning percentages of 45.66% and 38.31%, respectively, are not only significantly different from 50.00%, but they also suggest that a strategy of betting against Sagarin’s picks over these two time-frames would have been statistically significantly profitable. Though this evidence suggests that both experts’ projections lack any informational content for predicting winners against spreads, the possibility still exists that the experts’ seemingly impressive ability to pick straight-up winners will motivate sentimental betting activity."

Cat Grad · Post by **Cat Grad** » Thu Feb 17, 2005 12:15 pm

rtb: you're probably correct--read it someplace else too (I think).