Columbus, True Talent, and the Limits of Corsi
Good article from Tom Awad over at Hockey Prospectus, doing some nifty things with PDO and Corsi and finding important results concerning the performance we should expect from teams.
4 months ago
Bettman's Nightmare
17 comments
0 recs |
Email
PrintComments
I’d like to see this split up between save percentage and shooting percentage because if the shooting percentage part for Columbus is relatively minor we might chalk that up to luck while acknowledging that they have brutal goaltending bringing them down.
Writer for http://www.habseyesontheprize.com/
by Stephan Cooper on Jan 12, 2012 12:01 PM EST reply actions
For Columbus in particular, they’re -46 goals due to shooting and -73 goals due to goaltending. I don’t think the shooting can be considered to be “relatively minor”.
OK cool.
I’m also not quite sure if this is doing the right thing of combining powerplay and even strength play. I’m more sympathetic to the idea that there is sustainable shooting talent at the team level with the man-advantage than there is at even strength for one thing, the other thing is this could be skewed by relative amounts of time spent with the man advantage or with out it.
Going off timeonice for the past 3 seasons on 5 on 5 Columbus has scored 447 goals on 5579 shots for a percentage of 8.01 which is about average. So unless I did the math wrong it looks to me like the answer seems to me to be that Columbus has terrible finishing ability on special teams.
Writer for http://www.habseyesontheprize.com/
by Stephan Cooper on Jan 12, 2012 4:38 PM EST up reply actions
Boston has actually been the example I’ve used to intuitively dispute the notion of team-wide finishing talent. For 5-on-5 shooting percentage they’ve managed to be, going back in time starting with this year: first, fifth, dead last, second, 23rd. Only this year have they broken the pattern of wild swings in shooting percentages.
I think lack of PDO talent is far more likely than the ability to keep it high, if only because hanging on to bad goaltending can put it in the trash, but also because there is such a thing as a bad shooter. Goaltending talent makes some difference, but whether talents have the ability to push their shooting percentage high for long periods of time is another matter entirely.
Just a thought
What if your defencemen were the primary shooters in your offence? That would be another scenario where I could see PDO being artificially low.
"You can't polish a turd." -- George Carlin
Co-Manager at Arctic Ice Hockey
by Bettman's Nightmare on Jan 12, 2012 1:20 PM EST up reply actions
You’re right, although I doubt it could vary that much to justify it. Don’t let Gabe hear you: in effect you’re arguing for sustainable even-strength shot quality, which is a no-no!
Didn’t you mention as much, regarding Detroit, a while back?
Red Line Station and @RedArmyLine, featuring coverage of the most frustrating team in the NHL
I believe in next year.
by red army line on Jan 12, 2012 1:47 PM EST up reply actions
Fair point
Having made the comment, I’m going to look at the last few years and see if any teams have had a high percentage of their shots taken by defencemen (and, in general, if those percentages are volatile over time and across teams). I know Atlanta did it last year, but I suspect few teams have had a similar scenario.
"You can't polish a turd." -- George Carlin
Co-Manager at Arctic Ice Hockey
by Bettman's Nightmare on Jan 12, 2012 1:48 PM EST up reply actions
That would be my alter-ego, “Caricature Gabe.”
Defensemen shoot 13% worse than forwards, given the same shot locations:
http://www.puckprospectus.com/article.php?articleid=273
And 3rd and 4th liners shoot worse than 1st and 2nd liners (I should quantify that).
The issue is at the team-level, we don’t see massive differences in the ratio of shots taken by bad shooters vs good shooters.
Don’t worry, I know you understand the magnitudes involved probably better than anyone. You’re completely correct: we don’t see massive differences in the ratio of shots taken, which is why Ben’s explanation could only justify a sliver of shooting percentage differential. There’s no team where defensemen take 80% of the shots.
You can’t take one team as an example to dispute the notion of team-wide finishing talent. How about Vancouver, who’s been fifth, second, second and fifth? What you should be arguing (correctly) is that the deviation over one season is large enough that a high number in one season cannot be taken as evidence of finishing talent. What I’m showing here is that over multiple seasons you can tease it out.
Btw, how do we explain Boston crashing in the playoffs in 2008-09? Sunny Mehta made some serious money betting on PDO regression…
I’m not claiming that it was a proof of anything. I use it to make a point to more intuitive types, who often feel Boston’s offense is due to genuine finishing talent and aren’t particularly interested in mathematical arguments. It’s merely meant to be illustrative.
Comments For This Post Are Closed
Secondary Sidebar
User Tools
Related FanPosts on Arctic Ice Hockey
FanPosts
Community blog posts and discussion.
