How useful is Raw +/-?
by Hawerchuk on Oct 30, 2011 12:05 PM EDT
in Statistics
Email
Print
Someone asked friend of the blog Neil Greenberg how reliable individual +/- is. I think it's very reliable - in fact, you would do reasonably well signing the player who finishes dead last in the league in +/- every season provided he wasn't on an expansion team. Let's take a look at that list:
| Season | Player | GP | +/- | GP Year2 | +/- Year 2 |
| 2010-11 | Chris Phillips | 82 | -35 | ||
| 2009-10 | Patrick O'Sullivan | 73 | -35 | 31 | 4 |
| 2008-09 | Brendan Witt | 65 | -34 | 42 | -18 |
| 2007-08 | Radek Bonk | 79 | -31 | 66 | -12 |
| 2006-07 | RJ Umberger | 81 | -32 | 74 | 0 |
| 2005-06 | Mark Recchi | 83 | -36 | 82 | 1 |
| 2003-04 | Rico Fata |
73 | -46 | no season | no season |
| 2002-03 | Jay Bouwmeester | 82 | -29 | 61 | -15 |
| 2001-02 | Tyler Wright | 77 | -40 | 70 | -25 |
| 2000-01 | Patrice Brisebois | 77 | -31 | 71 | 9 |
| 1999-00 | Yannick Tremblay | 75 | -42 | 46 | -6 |
| 1998-99 | Darcy Tucker | 82 | -34 | 77 | -12 |
| 1997-98 | Paul Ysebaert | 82 | -43 | 10 | -5 |
| 1996-97 | Alexandre Daigle | 82 | -33 | 75 | -8 |
| 1995-96 | Owen Nolan | 81 | -33 | 72 | -19 |
| Craig Janney | 84 | -33 | 77 | -1 | |
| 1994-95 | Chris Dahlquist | 46 | -30 | 24 | -7 |
| 1993-94 | Gord Dineen | 77 | -52 | 9 | -5 |
| 1992-93 | Rob Zettler | 80 | -50 | 42 | -7 |
| Doug Zmolek | 84 | -50 | 68 | -9 | |
| Neil Wilkinson | 59 | -50 | 72 | 2 | |
| 1991-92 | Paul Fenton | 60 | -39 | retired | retired |
| 1990-91 | Michel Petit | 73 | -34 | 70 | -15 |
| 1989-90 | Bryan Fogarty | 45 | -47 | 45 | -11 |
| 1988-89 | Tom Fergus | 80 | -38 | 54 | -18 |
| Average | 74 | -38 | 56 | -8 |
More than half of these guys were in the middle of otherwise fine NHL careers and found themselves getting a lot of tough ice time on an awful team and, on top of that, had all of the bounces go against them. The top half of these players went from an average of 74 GP and -38 in their bad +/- season to 73 games and a -7 in their next season.
This relates back to our discussion of PDO on Friday. A lot of bad +/- is driven by a low on-ice shooting percentage and a low on-ice save percentage - PDO. And PDO has zero year-to-year correlation, so it's difficult to infer much skill from a bad +/-. And clearly a lot of decision-makers in the NHL do get that (otherwise they'd cut their +/- trailers) but I'm willing to bet that guys with poor +/- are still undervalued.
16 comments
|
2 recs |
Do you like this story?
Comments
I literally just fan posted on this but a quick quantitative summary is this; (RTM = Regression to the mean)
5GP 100% RTM
10GP 87.5% RTM
20GP-70GP 80-70% RTM
82G 60% RTM
>120GP 55% RTM
I totally agree with Gabe- It takes a lot of games before a player’s “true” +/- can faintly be seen, thus a bargain hunting GM might find some value in an over-valued stat.
I think it’s very reliable – in fact, you would do reasonably well signing the player who finishes dead last in the league in +/- every season provided he wasn’t on an expansion team.
Needs “not” or “don’t” somewhere.
Driving Play - The Blog with Three First Lines
I meant my strategy of signing these guys
by Hawerchuk on Oct 30, 2011 4:44 PM EDT via mobile up reply actions
whoosh, your point flew right over my head.
Driving Play - The Blog with Three First Lines
Heh, got it the second time. I’m slow but not that bad.
Driving Play - The Blog with Three First Lines
about weird +/- stats
http://www.nhl.com/ice/player.htm?id=8447531
The 74-75 season for Lynch when playing for Capitals. -54 in 20 games is one of the most amazing stats I’ve ever seen…
The fact that the 4 worst +/- seasons ever in NHL are from players of the 74-75 Capitals team (with Lynch playing a part of the season) says a lot of how terrible they were.
Brett Lebda: -3 in a 9-3 win
That has to be up there as one of the most hilariously amazing stats.
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
It’s quite impressive, now imagine him pulling that rating off for 19 more games… =)
(which seems very possible on the “right” team)
by humanindustry on Oct 31, 2011 9:15 PM EDT up reply actions
2003-04
Also, the worst +/- guy of the 2003-04 season was Rico Fata with -46 while playing for the Penguins.
Hehe, I went all the way to page 945 to check the all time worsts there, regardless of season, and noticed Fata by accident. I hate that the ones without a rating ends up after the ones with a minus rating…. =P
by humanindustry on Oct 31, 2011 9:14 PM EDT up reply actions
as a flames fan, I can tell you this is not surprising in the least
by Justin Azevedo on Nov 1, 2011 3:54 AM EDT up reply actions
Just like pitcher W-L, which could be thought of as a variation on +/-. Some pitchers are just randomly going to get a bunch of starts
in which the team doesn’t score runs or fields terribly behind him. Same with shifts. With 600 or whatever guys there are going to be a few who have terrible luck with respect to their teammates performance while they are on the ice. But if a guy keeps on being put out there he must not be that bad so the +/- leader is more likely to be unlucky than bad. Anyone who is really bad won’t get enough ice time to trail the league in +/-. Except that Lynch guy you mentioned. That’s bad! If a guy trails the league with only 20gp then he’s probably more bad than unlucky.
I guess the exception is guys who were really good but suddenly have become very bad. It takes a while for everyone to be convinced that the guy really has lost it and isn’t just unlucky or slumping so they keep playing him (think John Lackey).
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
