HFBoards (http://hfboards.hockeysfuture.com/index.php)
-   By The Numbers (http://hfboards.hockeysfuture.com/forumdisplay.php?f=241)
-   -   Predicting game outcomes (http://hfboards.hockeysfuture.com/showthread.php?t=1473949)

 imperfect 07-22-2013 10:07 PM

Predicting game outcomes

I haven't seen a lot of discussion about prediction of individual games on here or among the rest of the hockey stats community. All I've managed to find is this article from hockeyanalytics, and it's fairly basic. Does anyone know of other resources on this?

And a second question: is there any work done on how goal-making pace in the different periods are related to the full game? When I try to model total goals scored for a full game, poisson works fairly well but for specific periods it doesn't really fit. So does anyone know of any better way to model goal making in periods? Thanks in advance everyone!

 PavelDatsyuk 07-23-2013 05:47 PM

Quote:
 Originally Posted by imperfect (Post 69508875) I haven't seen a lot of discussion about prediction of individual games on here or among the rest of the hockey stats community. All I've managed to find is this article from hockeyanalytics, and it's fairly basic. Does anyone know of other resources on this? And a second question: is there any work done on how goal-making pace in the different periods are related to the full game? When I try to model total goals scored for a full game, poisson works fairly well but for specific periods it doesn't really fit. So does anyone know of any better way to model goal making in periods? Thanks in advance everyone!
Goal pace increases in the 2nd and third periods (third obviously due to empty net goals, 2nd for whatever reason I'm not entirely clear on still). I don't have the exact numbers or even really ballpark numbers to support but on the top online books, 2nd period scoring vs 1st period scoring is usually somewhere in the region of -130/+110 I believe. And third period vs 2nd period is usually a very, very slight favorite just going off memory.

^ of very little use I know, but perhaps you can make something of it

 SnarkSD 07-27-2013 05:47 PM

I did some analysis a year or so ago. You can find more about game-by-game probabilities here. I would not use goals for - against, but use shot based metrics, transformed to goals for better prediction.

 Timmer44 07-31-2013 02:48 PM

Quote:
 Originally Posted by PavelDatsyuk (Post 69538135) Goal pace increases in the 2nd and third periods (third obviously due to empty net goals, 2nd for whatever reason I'm not entirely clear on still). I don't have the exact numbers or even really ballpark numbers to support but on the top online books, 2nd period scoring vs 1st period scoring is usually somewhere in the region of -130/+110 I believe. And third period vs 2nd period is usually a very, very slight favorite just going off memory. ^ of very little use I know, but perhaps you can make something of it
Interesting. Maybe the long change?

 All times are GMT -5. The time now is 04:09 PM.