14 replies to this topic

[Triumph]

Want to know why Jay Pandolfo is one of the best defensive forwards in the game, why Vincent Lecavalier, not Sidney Crosby, was the true MVP of 2006-07, why Zach Parise was better than Jaromir Jagr, or why Scott Gomez was the 238th best player in the league last season?

http://www.hockeyanalytics.com

Here's a link to the Excel file where the players are rated: http://www.hockeyana...files/PC_07.xls - On the Devils, Colin White is surprisingly low, but probably part of that is his total offensive inability and propensity to take penalties. Patrik Elias was also rated barely above Jamie Langenbrunner.

Edited by Triumph, 06 November 2007 - 04:05 PM.

[SteveStevens]

I dled the excel, pretty awesome!

[SteveStevens]

Paul Martin is #99, which is pretty impressive,and thissystem works because of how he grades everything

Parise at 52 is impressive also.

Also funny to see zajac is 237 and gomez is 238

[halfsharkalligatorhalfman]

interesting but again, doesn't work on the same level as baseball sabremetrics where there's thousands of one on one battles (pitcher v. hitter) in a season since hockey is such a team game.

[Triumph]

interesting but again, doesn't work on the same level as baseball sabremetrics where there's thousands of one on one battles (pitcher v. hitter) in a season since hockey is such a team game.

No, and the author admits that. I think there's plenty of flaws with it, but it's in a very early stage right now - it took baseball sabermetrics 20 years, and baseball is a mathematizable game. Hockey isn't.

Interesting how Zubrus only rates a 36 over the entire season - which puts him around 320th or so overall.

[SteveStevens]

It is flawed but its good to see in a direct way, if you take it watered down its pretty good.

[halfsharkalligatorhalfman]

i guess but its way too inaccurate as a means to compare players and get an appropriate measurement of their relative contribution, especially for players on different teams. Its just a good way to confirm past prejudices and observations (like that dan cloutier sucks)

[nmadd]

Thanks for the link. Stats are one of the big reasons why I like baseball so much. Even though this isn't quite the same, it's good to see something like this for the NHL.

Its just a good way to confirm past prejudices and observations (like that dan cloutier sucks)

...and that Marty is a stud.

[Triumph]

i guess but its way too inaccurate as a means to compare players and get an appropriate measurement of their relative contribution, especially for players on different teams. Its just a good way to confirm past prejudices and observations (like that dan cloutier sucks)

It's a benchmark. I think intuition will always play a huge role in evaluating hockey players - and this system has no way of rating physical play, forechecking, etc. It does seem to do a very good job of showing that certain players are highly overrated (Gomez, Malik) and certain ones are undervalued (Rachunek was only 4 points below Malik in terms of total performance last season).

There's no real way of isolating all the noise in hockey - coming up with something like a 'replacement player' in baseball seems an impossible task. Still, while hockey can never get close to baseball in terms of numbers, it can certainly help us evaluate just why the GM in Phoenix is a moron.

[Devils731]

I don't like numbers! I prefer more ambiguous things. </sarcasm>

[hockeymetrics]

i have to look it over more because i didn't notice what the ratings are totally based on, but offensive statistics have to be put into perspective.
Lecavalier's stats might be better because he scored 16 more goals, but he's in a better situation being used to the best of his ability. His best skill is goal scoring and he had 4 excellent playmakers setting him up - St. Louis, Richards, Boyle, and Prospal. The Penguins, meanwhile, should be ashamed they can't find a legit wing to pair with Crosby. That he could manage 84 assists without a pure goal scorer is testament to his jaw dropping, unbelievable talent. No doubt the Penguins have loads of talent, but not the right talent to best make use of Crosby. The worst part is (for fans who want to see him shine as bright as he can) the type of player he needs is extremely plentiful - Ryder, Pominville, Gionta, Kotalik, Lupul, Carter, Dawes, Callahan, Svatos, Tambellini, Jason Blake, etc.

[Triumph]

i have to look it over more because i didn't notice what the ratings are totally based on, but offensive statistics have to be put into perspective.
Lecavalier's stats might be better because he scored 16 more goals, but he's in a better situation being used to the best of his ability. His best skill is goal scoring and he had 4 excellent playmakers setting him up - St. Louis, Richards, Boyle, and Prospal. The Penguins, meanwhile, should be ashamed they can't find a legit wing to pair with Crosby. That he could manage 84 assists without a pure goal scorer is testament to his jaw dropping, unbelievable talent. No doubt the Penguins have loads of talent, but not the right talent to best make use of Crosby. The worst part is (for fans who want to see him shine as bright as he can) the type of player he needs is extremely plentiful - Ryder, Pominville, Gionta, Kotalik, Lupul, Carter, Dawes, Callahan, Svatos, Tambellini, Jason Blake, etc.

I don't think he revealed his methodology in full - much as the formulae for VORP, EqA, etc. in baseball and DVOA and DPAR in football are proprietary.

It is embarassing that the Penguins dress three players on a nightly basis who barely belong in the NHL, then stick them on their top 3 lines. Ryan Malone, Mark Recchi, and Gary Roberts are not very good players, at all. By the same token, though, the Penguins should take advantage of Crosby's wonderful abilities by NOT spending a lot on a guy who plays beside him. Sykora is a perfect low-cost option who could score 30-35 there this season easily. Paying top-dollar for a guy to score 50 with Crosby seems counterproductive - anyone not signed long-term who plays next to Crosby, his contract will explode, and it's tough to justify signing a mediocre scorer like Sykora very long-term.

Edited by Triumph, 08 November 2007 - 12:54 PM.

[harpua19]

Nice find some interesting stuff

[halfsharkalligatorhalfman]

the fact that it says rachunek is anywhere near as good as Malik shows how completely worthless it is.

Just kidding. It's interesting and fun to look through. I'd never make a judgement on a player based on it, but I may look out for certain players on my own because of it.

[Triumph]

the fact that it says rachunek is anywhere near as good as Malik shows how completely worthless it is.

Just kidding. It's interesting and fun to look through. I'd never make a judgement on a player based on it, but I may look out for certain players on my own because of it.

actually I think that's part of its merit. Its problem seems to be overemphasizing disciplined players - Chris Neil receives a minus for last season and Avery is barely a plus as a Ranger. I wonder if it takes into account coincedental minors. I do think that goes overlooked, though - a penalty is not only bad because teams score on PP's a certain % of the time but your team is very unlikely to score during those two minutes.

Edited by Triumph, 08 November 2007 - 05:05 PM.