Linez

gtphd said:
$50,000 on GT +13
Click to expand...

Think this was meant for the prediction contest.

The line isn't that different from Sagarin Predictor +3 for home advantage. I guess it would be GT +7 if it were at BDS, which is how the game actually finished last year.

I stopped trying to predict the future as far as football a long time ago, unless it was for multiple games. There's just too much inherent variance for a single game. In Moneyball, Billy Beane notes that he stopped caring about 5-game wild card series because it is basically a crapshoot. The variance in baseball is much clearer than football, since the line between hits and outs can often come down to luck.

But the variance in football is staggering any way you look at it. If you take Vegas lines as unbiased predictors of a future game, the standard deviation of a game versus that predictor is 12 points for the NFL and 15 points for college.

So, if we played UGA 100 times and Vegas is correct about +13 for the average game, the results would form a normal distribution with an SD of 15 points. So the result would be between -2 and +28 GT. 16% of the time the result would be better than -2 GT and 16% it would be worse than +28 GT. That's pretty incredible when you think about it.

So yeah, whatever. öööö Georgia.
 
floridajacket said:
Think this was meant for the prediction contest.

The line isn't that different from Sagarin Predictor +3 for home advantage. I guess it would be GT +7 if it were at BDS, which is how the game actually finished last year.

I stopped trying to predict the future as far as football a long time ago, unless it was for multiple games. There's just too much inherent variance for a single game. In Moneyball, Billy Beane notes that he stopped caring about 5-game wild card series because it is basically a crapshoot. The variance in baseball is much clearer than football, since the line between hits and outs can often come down to luck.

But the variance in football is staggering any way you look at it. If you take Vegas lines as unbiased predictors of a future game, the standard deviation of a game versus that predictor is 12 points for the NFL and 15 points for college.

So, if we played UGA 100 times and Vegas is correct about +13 for the average game, the results would form a normal distribution with an SD of 15 points. So the result would be between -2 and +28 GT. 16% of the time the result would be better than -2 GT and 16% it would be worse than +28 GT. That's pretty incredible when you think about it.

So yeah, whatever. öööö Georgia.
Click to expand...
this is how floridajacket trash talks, beats you to death with standard deviation
 
I'm running some different normality tests in R on previous college football data while Uga is dying a horrible death in my basement.
 
You must log in or register to reply here.
Back
Top