View Single Post
Old 12-05-2014, 08:14 PM   #200
Flames Fan, Ph.D.
#1 Goaltender
 
Flames Fan, Ph.D.'s Avatar
 
Join Date: Mar 2006
Location: Underground
Exp:
Default

Quote:
Originally Posted by Inglewood Jack View Post
any good stats cruncher will readily admit the current metrics are very rough. once they get that system in place that tracks every player and puck movement, that will be a quantum leap forward and we'll really start seeing solid analytics.

the most annoying thing about the stats hijackers will be the broken clock/blind squirrel effect. when the Flames inevitably go on a losing streak (and they will), those guys will get even louder and more obnoxious, like "see? I was right all along"...ignoring the fact that they looked totally clueless for several months prior.
From what I can find, 5 year correlations between Fenwick and Corsi For versus wins is 0.25 - 0.3ish. If this is true, I can see why people are annoyed hearing about them to describe / predict team performances... because those correlation values basically say no correlation.

If I submitted a paper claiming causative correlations based on values of 0.25-0.3, I'd get summarily dismissed.
Flames Fan, Ph.D. is offline   Reply With Quote
The Following User Says Thank You to Flames Fan, Ph.D. For This Useful Post: