Monday, October 22, 2007

Behind The Net is back

Mirtle blog approved™ stats man Gabe Desjardins has pumped out his first set of numbers for the 2007-08 season, and he says he's planning on having daily updates throughout the year.

Lucky us.

It's early, but you can see players like Nik Antropov, Kyle Brodziak, Bill Guerin and Pierre-Marc Bouchard have been pretty productive without having many even-strength goals scored when they're on the ice. On the flip side, it's been a tough start for Michael Nylander, Jarret Stolll, Dainius Zubrus and, yes, even Bryan McCabe.

Looking at the numbers, it makes you wonder just why Paul Maurice wouldn't put Antropov and McCabe out on the ice at the same time.


At 2:46 a.m., October 22, 2007, Anonymous Anonymous said...

There's 10-12 players on ice at any given (even strenght) time. To calculate numbers on individual basis lacks in credibility - big time.

Imagine having same kind of "idea" in science. 90% unknown, 10% fact...

At 9:00 a.m., October 22, 2007, Blogger Adam C said...

Yeah, imagine scientists trying to learn about a subject by comparing it to a control group.

Next time, follow the link.

At 9:51 a.m., October 22, 2007, Anonymous Anonymous said...

Don't these numbers only become useful when the Linemate Quality and Opposition Quality adjustments have been made?

Where can I find a discussion of the limitations of the behindthenet calculations?

At 11:06 a.m., October 22, 2007, Blogger saskhab said...

Early returns aren't really credible for most cases anyways. The site isn't the be all and end all, but it does show general effectiveness pretty well.

At 12:15 p.m., October 22, 2007, Blogger Darrell said...

First, this rating system clearly penalizes players for being on good teams.

Second, it clearly needs an adjustment for opposing player quality. I am only going to do Ottawa, but the results for them are stagering.

Ottawa's rankings are as followsm with only the following players being "even" or better:

Antoine Vermette
Chris Kelly
Luke Richardson
Chris Phillips
Chris Neil
Josh Henessey
Anton Volchenkov
Joe Corvo

That leaves the entire Pizza line out in the cold, despite them being 4th, 6ht and 8th in league scoring, and regularly dominating play.

At 12:17 p.m., October 22, 2007, Blogger James Mirtle said...

That line has just been on the ice for more goals against so far — it's pretty early to be drawing any significant conclusions from the data.

At 2:45 p.m., October 22, 2007, Anonymous Anonymous said...

That list of Ottawa players proves everything. Same as plus/minus proved last season that Tom Preissing was Ottawa's best defenceman.

Too bad Preissing and Corvo didn't impress their coach who chose to keep them on the bench quite frequently.

Next time I will follow the link so I can find out all of this.

At 2:54 p.m., October 22, 2007, Anonymous Frank said...

I put together the list for Vancouver - the team I know best - and it was right on the mark. The five worst ranked players were:
(1) Shannon (C) -8.7
(2) Bieksa (D) -3.0
(3) Naslund (LW) -2.8
(4) Ritchie (C) -2.5
(5) Krajicek (D) -2.4
Anyone watching the games didn't neeed the stats to tell us this, but they certainly help to bolster our subjective views. On the surprise side, Weaver is ranked as our second best d-man, while most fans believe he has been our worst d - man, next to Bieksa.

Thankfully, Shannon is back in the AHL learning to backcheck (thaks a lot Brian Burke for the trade) - and I'm certain Bieksa will play his way out of this slump.

However, it looks like we will have to put up with another (but final) year of ineptitude from our $6 million captain - Naslund. Perhaps we can trade him for McCabe!

Also, the Naslund experience should really put fear in the hearts of GMs who are signing players to 6 or 8 year deals. Thankfully Naslund was signed for only 3 years and we will be rid of him after this year. Can you imagine if he was signed to a 6 or 8 year deal!!

At 2:58 p.m., October 22, 2007, Blogger James Mirtle said...

Heh, I'll never understand why people who apparently dislike everything I post come back continually to tell me so.

No matter.

Look, there is clear, objective data available on Behind The Net that you can't get anywhere else: The on ice goals for and goals against figures are terrific reference points for how players are playing at even strength, even without having strength of competition/linemates data.

Is it better to have those other metrics? Sure. But we don't use qualifiers like that in determining the Art Ross winner, or the Norris, etc.

Players on better teams, with better goaltenders, are simply going to get more accolades and have better statistics in a whole variety of ways. That doesn't make said statistics useless.

At 3:05 p.m., October 22, 2007, Blogger Adam C said...

It's a more reflective stat than +/-, and that gets pretty prominent play. At least this way you can compare players from different teams.

I've been thinking Naslund for McCabe could work out for both teams, but Vancouver would be stuck with the extra years on Bryan's contract.

At 4:19 p.m., October 22, 2007, Anonymous Rod said...

Maybe I'm just too tired, but could somebody please explain how these add up?...

GFON/60 5.6
GAON/60 0.5
PMON/60 4.0???
GFOFF/60 2.2
GAOFF/60 3.2
PMOFF/60 -1.0

Is the PMON/60 column supposed to be the difference between GFON/60 and GAON/60?
- If so, the PMON/60 column doesn't add up for any player with a non-zero GAON/60.
- If it's not supposed to be the difference, how is the PMON/60 column arrived at? PMOFF/60 is a simple difference, so why not PMON/60?

Again, any non-zero GAON/60 in the tables can be used to see this:
GFON/60 6.7
GAON/60 1.3
PMON/60 2.9???
GFOFF/60 2.9
GAOFF/60 4.2
PMOFF/60 -1.3

Clarification would definitely be appreciated. Thanks.

At 4:26 p.m., October 22, 2007, Blogger James Mirtle said...

Your best bet would be to email Gabe; apparently they've changed the boxscore system this year and there've been some bugs.

I honestly haven't had time to go through the data that closely yet.

At 4:43 p.m., October 22, 2007, Blogger Hawerchuk said...

Looks like the GAON/60 column is the wrong data, though the rest are all correct. I rushed through this a bit to post a table yesterday. Anyways, I'll check it out and fix it tonight.

At 10:02 p.m., October 22, 2007, Blogger Hawerchuk said...

ok, GAON/60 is fixed. Let the lively debate over the significance of these numbers begin...

At 10:14 a.m., October 23, 2007, Blogger stevied said...

It looks to me like this system rewards good players with mediocre teammates and penalizes players on teams with more depth - since it is comparing one player against every other player for the team. It is rather early in the year so it would be interesting to see the results of this with more data.

At 10:17 a.m., October 23, 2007, Blogger stevied said...

Also - shouldn't it be normalized by minutes played per game?

At 6:28 p.m., October 23, 2007, Blogger Hawerchuk said...

1) See last year's numbers if you need higher statistical significance.

2) The statistics are presented per 60 minutes of ice time. These are rate stats, so everybody is on the same footing.

3) You're free to argue about what the numbers represent. Remember - goals, assists and +/- reward players with good teammates.


Post a Comment

<< Home

Links to this post:

Create a Link


Free Page Rank Checker
eXTReMe Tracker