View Single Post
02-27-2013, 04:12 PM
Doing Nothing
Jarick's Avatar
Join Date: Jul 2007
Location: St Paul, MN
Country: United States
Posts: 25,251
vCash: 500

Declining value of Fenwick and Corsi with increased sample size

The last several days I have been playing around a fair bit with team data and analyzing various metrics for their usefulness in predicting future outcomes and I have come across some interesting observations. Specifically, with more years of data, fenwick becomes significantly less important/valuable while goals and the percentages become more important/valuable.
He shows 87% correlation between goal differential and on-ice success.

Welcome to 2011

I haven't delved into it much, but when I dipped my toe into it, the "Pythagorean Equation" [GF^2 / (GF^2 + GA^2)] had an 84.7% r-squared while looking at shot differential was 21.3%. This is going back to 07-08. I will go further when I've got the time.
And again

Quote: lockout I've got 211 team seasons with 15,628 games played, 44,114 goals, and 467,707 shots.

It's just that not all shots are equal. If all systems and goalies were equal, you could look at shots. But they aren't.

I just don't understand how they don't look at whether or not the stat is a good predictor or not. They have theory after theory after theory that doesn't bear out.

Jarick is offline