Win Probability in Hockey

[TheChamber91]

Does anyone have a good formula for calculating win probability in hockey? One that incorporates time left in the game, winning percentage, etc.

Click to read more...

 

[Doctor No]

I've put together a few models using a Bayesian framework, but I'm not particularly satisfied with them just yet.

 

[LeHab]

I've put together a few models using a Bayesian framework, but I'm not particularly satisfied with them just yet.

Is that with a single or multiple variables? Which one? Having a dynamic win chance predictor which factors things like time left, current score, shots etc... could be quite interesting.

 

[TheChamber91]

Is that with a single or multiple variables? Which one? Having a dynamic win chance predictor which factors things like time left, current score, shots etc... could be quite interesting.

I'm interested too, can we see the unfinished models? I need a good model for a study I want to begin.

 

[Hockey Outsider]

Does anyone have a good formula for calculating win probability in hockey? One that incorporates time left in the game, winning percentage, etc.

Here's an interesting essay on this topic - http://www.hockeyanalytics.com/Research_files/Poisson_Toolbox.pdf.

The application to a single game (ie what are the odds of a team winning based on a lead/deficit of X, with Y minutes remaining) starts on page 30.