View Single Post
Old
11-18-2012, 10:58 AM
  #10
Sureves
Registered User
 
Sureves's Avatar
 
Join Date: Sep 2008
Location: Ottawa
Posts: 9,184
vCash: 500
Quote:
Originally Posted by eklunds source View Post
I'm a fan of anyone who uses hockeyanalysis, but mentioning goals.. Well, they're sort of irrelevant.

Goals are obviously what win or lose you hockey games, but they're so unpredictable. Perry directed a pretty similar number of pucks on net this year and last (13 shots less in 2 less games) and scored 13 goals less. Kuba had an almost identical shot rate for/against in his last two seasons in Ottawa, yet went from -26 to +26.

Over a large enough sample size, those things even out, but one season isn't large enough. That's why you have players who have years like Getzlaf - who looks awful statistically but really just went through two ends of the on-ice shooting percentage spectrum..

Shot rates are much more repeatable, thus much more accurate and a much better predicter of future success. Talking about goals will just lead to people thinking "so he creates higher quality chances", which might happen... But if it does, the effect is so small that it gets buried in statistical noise..



Edit: more detailed analysis of Suter and Weber: http://nhlnumbers.com/2012/7/18/acco...i-rel-and-wowy

They only played about 3 games without each other. Not nearly enough of a sample size to say for certain how they influence the game apart from their partner.. As the article says, any team in the league can look amazing or awful over a 3 game stretch.
Agreed on the goals topic, Corsi is a better representation for this type of analysis.

Sureves is online now   Reply With Quote