View Single Post
Old
03-30-2011, 01:39 AM
  #63
tpb209
Registered User
 
Join Date: Mar 2006
Location: NJ
Country: United States
Posts: 444
vCash: 500
Quote:
Originally Posted by Tawnos View Post
It's really annoying when people discredit the +/- statistic.

Yes, you're right. There is no way to universally apply the stat in the same way to every player league wide. That doesn't mean it's a bad circumstantial stat. In fact, I'd say it's a very good circumstantial stat. Marek Malik's +/- was high because of Jagr and co. I don't think he was as bad as most people made him out to be, but that stat was certainly inflated and our observations made light of that fact. On the other hand, DelZotto's +/- last year was horrible. We might look at this case of the stat being an affirmation of what we're seeing with our eyes: that DelZotto was playing pretty terrible defense.

In regards to the topic: people say that +/- for defensemen gets inflated by good offensive forwards playing well. But I have to ask you, is that the case here? I mean, can we reasonably point out that this pair does a VERY good job on the ice defensively, albeit against 2nd and 3rd lines more often that 1st lines? We see it with our own eyes and the statistic backs it up. we want cup says that our team spends a lot of time cycling the puck around in the offensive zone. In the Rangers scheme, the D play such a large part of that... you can't say the forwards are inflating the stat in isolation. The defense are inflating their own stat... imagine that! It's an indicator of how good a job the D are doing!!!

They aren't playing against the opponent's absolute best and they aren't having their numbers inflated by offense. Their +/- still tells me that they're doing a damn good job out there.

The lack of understanding of how to apply this statistic to reality is mind-boggling. There is a reason this is kept track of, despite the wide attempts by fans to discredit it.
The problem with the +/- stat is that a person could have absolutely nothing to do with the goal for. The player could also have absolutely nothing to do with the goal against.

The supposition is that these outliers will even out over the long term, but for many players this may not happen. It is the same way for many stats. This is the main problem people have with traditional baseball stats. A pitcher could give up 8 runs in 5 innings and still get a win. Does that "win" mean that this pitcher is better than a pitcher with one less win? No, it does not. In my opinion, logical, rational statistics only represent what a specific player is responsible for. It really is not possible, at this point, to do that in overall hockey statistics. There is too much white noise. In baseball, it is easier to create advanced metrics because almost everything is made up of 1v1 matchups.

The problem is, without watching every game, a person does not know how well a player is playing. Ranger fans, for the most part, know that McD and Sauer are playing well, but they may or may not be playing as well as their +/- indicates.

(If this isn't clear or well thought out, I apologize. It was trivia night at the bar and I had a few.)

tpb209 is offline   Reply With Quote