View Single Post
Old
08-19-2012, 11:36 PM
  #32
Ohashi_Jouzu
Registered User
 
Ohashi_Jouzu's Avatar
 
Join Date: Apr 2007
Location: Halifax
Country: Japan
Posts: 20,826
vCash: 500
Quote:
Originally Posted by tonynyy View Post
The thing about shot differential stats like Corsi is that they tend to underrate players that play tough defensive minutes. The shutdown forwards of the league play against the other team's best nightly, so their Corsi ratings tend to be low. Consequently, players who are subject to heavy line-matching, such as Desharnais, tend to have lower QoC numbers, even though they were oftentimes matched up against the opposition's best defensive players.
I linked an article earlier from nhlnumbers.com, which advances the theory that we should measure quality of competition by looking at a player's opposition's average ice time. If a forward is getting 18 minutes of ice time a game, he's likely to be one of the best on his team. The article (linked below too) shows graphs for the each team, showing a player's average opposing defenseman ice time and forward ice time. IMO, it gives a really good picture of a player's usage and how other teams tend to match up against him. For example, Ovechkin's opposing defensemen tended to play a lot of minutes, but not the forwards. This means that he faced the league's best defensemen last year, but not it's top scoring threats. On the other hand, a guy like Brandon Sutter played some pretty weak competition in terms of opposing defensemen, but faced the other team's best forwards night in night out. A guy like Datsyuk faced both the best defensemen and the best forwards, which makes a lot of sense since he's widely perceived as one of the best two-way forwards in the NHL.
My point isn't that Corsi isn't a good measure of QoC. It does show how well a player's opponents drive possession. However, it tends to be seen as a be-all end-all of hockey stats in a lot of circles, and doesn't get taken in proper context enough, IMO.

Link: http://nhlnumbers.com/2012/8/17/time...r-all-30-teams
Well, in my opinion it's incredibly short-sighted to look exclusively at ice time, since I imagine you'd end up with an extremely wide range of talent/productivity/defensive acumen if you just sorted the NHL by icetime and started grouping guys who are close together. The icetime thing is interesting, but not as important as measuring how productive players are with the role/icetime they fill, and THEN looking back at who matched up against who. Comparing the resulting charts of the6thsens and nhlnumbers, I personally think the6thsens gets closest to "categorizing" the players "accurately" based on players' roles AND how they did in them. Also, I don't think it's necessarily beneficial to divide the focus between defensemen and forwards faced, since players (and their linemates) must be productive against 5 man units across 3 zones of ice - not just a pair of defensemen in one end of the ice.

And furthermore, I mean, if you think Corsi stats underrate defensive players, you must be floored by how favourably Eller still ranks on those lists versus Desharnais. After all, Eller held down far more defensive responsibilities and yet still came out on top by far in the categories I mentioned earlier.


Last edited by Ohashi_Jouzu: 08-19-2012 at 11:49 PM.
Ohashi_Jouzu is offline   Reply With Quote