More To The Sam Gagner vs. Chosen Line Comparisons
BSOD has everyone in Edmonton in love with all players drafted since 2008 while simultaneously crapping all over every player drafted prior to 2008. But when those players are compared directly to each other, a noticeable trend (at least by the possession metrics) is evident. The shiny baubles (except for Taylor Hall) aren't outpacing the musty old things.
The horizontal axis shows qualcomp, specifically Corsi relative quality of competition taken from the venerable and terrifying Gabriel Desjardins' behindthenet.ca. The vertical axis shows percentage of faceoffs taken in the defensive zone, again from the venerable and terrifying one. The bubbles are color-coded: blue means the player in the bubble has a positive zonestart-adjusted Corsi, red a negative. Finally, the size of the bubble indicates absolute value zonestart-adjusted Corsi. The shapes are matched by age, so direct comparisons can be made between the prospects everyone should like more (Ryan Nugent-Hopkins, Jordan Eberle) and the prospect it's okay to hate (Sam Gagner).
*Click to enlarge or find the full-sized chart here.
12 comments
|
0 recs |
Do you like this story?
Comments
I’m a little confused. In the Gagner article you have one of these charts and his 11-12 season is a small, positive blue dot, but in this chart his 22 year old season (which would be this season, right?) is a red, negative dot. What am I missing?
by melancholyculkin on Feb 19, 2012 10:44 AM MST reply actions
Couple of weeks between the data snapshots, I imagine.
Editor of The Copper & Blue, and leader of The Cult Of Hartikainen.
What doesn’t make sense is how Gagner’s bubble has gone from positive to negative during a time he’s been killing it with Hall and Eberle. Makes me wonder just how punitive is that adjustment coefficient in Zone Star Adjusted Corsi.
Oilers fan through thick, thin and anorexic. Writer for The Cult of Hockey.
by Bruce McCurdy on Feb 19, 2012 12:02 PM MST up reply actions
They really aren’t killing it by possession though.
Editor of The Copper & Blue, and leader of The Cult Of Hartikainen.
Doesn’t this miss the point.
It would seem that the weighted average used to adjust for zone starts could easily distort the data when a player has few starts in a particualr zone. For this reason it is perfectly reasonable to question whether the conclussion you draw is at all valid.
How is it distorting the data? As I pointed out here:
http://www.coppernblue.com/2012/2/12/959577/horcoff-hemsky-vs-gagner-eberle
The team was relying on goaltending. They were all getting beaten at evens. The Chicago game alone sent their numbers deep into the red.
Editor of The Copper & Blue, and leader of The Cult Of Hartikainen.
I think you have again missed the point. The issue is the nature of the statistic itself which is a weighted average. If one of the factors is a rather rare event for one player vs another the weighting can create a significant distortion. In essence it is similar to what happens if you try and use statistics such as XXX/60 with players who play only a few minutes per game over a small number of games.
Bruce McCurdy asked if it was possible that the adjustment method is punitive. You did not answer this but rather first reiterated the conclussion and now you reference the same stat beening used over a different period oftime.
If you want to make definitive statements using a formula like this isn’t it important to at least acknowledge the potential flaws in the analysis.
They’re under water without the adjustment, that’s my point. If they were in the black without the ZS formula, I might find the question interesting.
Editor of The Copper & Blue, and leader of The Cult Of Hartikainen.
I had the same basic question, but when I looked at players who had similar ZS assignments, the group was, on average, in the black by adjusted Corsi.
The biggest fanana of the Havana Bananas.
by Scott Reynolds on Feb 20, 2012 9:58 AM MST up reply actions

by 


























