View Single Post
Old
02-19-2012, 04:08 PM
  #35
Damaged Goods
Registered User
 
Damaged Goods's Avatar
 
Join Date: Feb 2009
Location: Philadelphia
Country: United States
Posts: 2,026
vCash: 500
Quote:
Originally Posted by sa cyred View Post
This is exactly what I am saying... where are we disagreeing? Is that considered luck or statistical contributions from other members of the team? Also how do you classify his skill? His statistics show that he was average defensively while contributing offensively.
That's the problem with taking these numbers at face value. You need to apply some Bayesian theory. Which is more likely: Gus right now is still only marginal player who happened to get lucky with his +/- numbers in a small sample size, or Gus just stepped onto the NHL scene with the rare kind of skills needed to substantially tip SH% and SV% at both ends of the ice... ? I would bet on the former.

Quote:
Originally Posted by sa cyred View Post
When looking at hockey advanced stats, you need to take in almost everything to get a clear view of how a player is doing. This includes QOC, QOT, Corsi, Fenwick, etc etc. If we start qualifying these as luck, than how/what do we base a player off of? Just watching them? I see that as more opinionated than actual fact.
"Viewing comprehension" still plays a gigantic role because NHL stats are nowhere near advanced enough to deal with the randomness issue very well on their own. These on-ice stats don't even attempt to address it, and I assume any serious champion of the stat would caution you from utilizing it they way you are. I don't think it's meant to draw the kind of inferences you want to draw from it.

Damaged Goods is offline   Reply With Quote