I’d also toss in GWG — it’s never been meaningful to me.

I wouldn’t junk GAA — I’d just make people take a seminar on what it means before they’re ever allowed to use it in a sentence. Save% tells me about the goalie. GAA — after seeing save% — gives me a quick, illustrative reference to what the goalie’s dealing with night after night. For me, the teammates and opposition are implied within it.

So much of these stat category arguments fall on the simplicity/accessibility of them. People just want to be able to scan a sheet, recognize a stat, and feel they learned something quickly. If there were a quick, established acronym that people knew for some of the reports James runs, they’d probably be accepted pretty quickly (accept for by grumpy reporters, I guess). We need a Global Panel of Hockey Stats dictating these changes. Actually, given how long it took baseball statheads to gain even today’s level acceptance, we probably need to storm the league office and those of its broadcast partners.

yep. It’s more a barometer of offense than defense. No need to junk it entirely, but let’s not use it to evaluate a player’s defensive ability

Surely you can see how plus-minus is more problematic than goals scored.

huh? what’d you steal?

if ERA were calculated with unearned runs factored in.

That’s why the Baseball Prospectus guys (among others) have come up with ways to “normalize” ERA based on defense, park, league, you name it.

It’s the next logical step – applying the same kind of analysis to hockey. That’s why I love the work of people like The Forechecker.

I’ve never understood the usefulness of the +/- stat. It reminds me of the “Runs” stat in baseball. Total junk.

This post receives my full endorsement.

His +46 (and the Selke votes) in 06-07 was a result of playing against inferior competition while the Briere and Drury lines soaked up the top defenders.

It’s one of the reasons why he struggled so mightily the first half of 07-08 – the adjustment period going from playing against 3rd and 4th lines and 3rd defense pairings to playing against “real” checking lines and top defense pairs.

“If a player’s +/- changes while his linemates stay the same, that says something about his play.”

Is that even possible? I mean, a + or – is assigned to everyone on the ice, irrespective of their relative roles in the event in question. If one person’s +/- changes while his "linemates"’ didn’t, that means he doesn’t have the same linemates anymore (or he’s getting nailed for a series of SHGA).

“Maybe we can just get people to stop basing their player evaluations (like ridiculous Selke votes) on a clearly flawed stat.”

I agree wholeheartedly, but good luck with that. That’d be like teaching a TV analyst to ask a good question between periods. ;)

That was a much better job of what I was trying to say in my little one liner.

Even something as simple as what I do with the best defensive defencemen and forwards lists is a whole lot more valuable than plus-minus. A lot of the players I’ve got highlighted up at the top are terrific defensively but post average plus-minus stats.

Justin – it turns out that baseball needs to be analyzed as a continuous game. The only thing that really lent itself to discrete analysis was plate appearance outcomes for hitters. No serious analyst really uses the statistics that came out of discrete analysis anymore. The pitchf/x system tracks pitch trajectories so we can analyze batter and pitcher strengths and weaknesses. And teams use a large number of observers to track fielding and positioning.

Hockey’s already past the discrete event phase – look at how many sites track head-to-head icetime? Or track how groups of players do when they play together? Even analyzing shooting or save percentage by distance and location arguably represents continuous dynamic statistical analysis.

How about if you go down into the career minuses? Kirk Muller has a career plus-minus of -146!

Randy Carlyle, -135
Bernie Federko, -132
Wendel Clark, -129

etc., etc.

Terrific Defensively . Top 15 in goals scored. 12th in career points/game

+/- of +30.

What I’d really like to see is some of those other stats come into play more often. I called plus-minus a crutch because it’s often used like it’s the be all and end all of measuring players’ performance. Way too often.

If I stop using it altogether, I bet I don’t lose much analytical ability. We’ll find out.

Right here big guy.

sort of true, but I think your comment is misleading. Traditional plus-minus is a number, like plus-seventeen. Isn’t Desjardins using a per-60-minute rate?

It’s all going to be flawed with data collection as it is, but I don’t mind Desjardins moving in the right direction. I find his numbers to be generally useful (you can identify players based on roles, which I think helps a ton).

I think you’re missing what Desjardins’ quality of competition is based on entirely. It’s opponents’ goals for when on the ice, goals against when on the ice, goals for when off the ice and goals against when off the ice.

That’s significantly different than simply measuring goals for and against for a particular player and keeping in shorthanded goals, 4-on-4, etc. It’s measuring the figure for every single player on both teams and adjusting for how their numbers are different from teammates/opponents.

hi David – please email me at info@behindthenet.ca with your questions about qualcomp. I’d be happy to answer them.

He’s using an altered form of “traditional plus/minus” as a part of the equation, but one that’s weighted by considering what happens when players aren’t on the ice as well. It’s pretty nifty.

But I’m not sure that it matters if it’s finding are stated as a seasonal rate, as plus/minus is, or on a per 60 minute ES rate.

It matters to me. If two guys were plus-thirty and minus-twenty over fifty games, but one played twice the ice time of the other, I wouldn’t want them considered identical for these purposes.

I swear by his qualcomp concept, but the numbers, well, not so much . . .

I think that’s where a lot of people sit. I don’t pretend that it’s a flawless ranking, but it separates the wheat from the chaff.

Stats have to appeal to the casual fan, not just the math geeks.

Baloney. Stats don’t have to appeal to anyone, I don’t think. If it’s not your cup of tea, don’t drink it. But I don’t think the key to determining a statistic’s usefulness is its popularity.

Rates are so basic and simple. That’s the thing — all they do is count the number of goals for and against and divide by ice time.

You need a slide rule for that? Honestly?

a stat like Desjardin’s adjusted Plus/Minus, which uses as its data set the highly lamented Plus/Minus, is good?

If you don’t understand what a stat is, try not to describe it as an “abortion”.

Francis: Over-rated stats accumulater exposed by Plus-Minus.

This is getting ridiculous. People who have clearly never seen Francis play would rather base their opinion on an easily-discredited stat than on the opinions of people who actually watched, played with, or coached him.

Let’s look at some other Selke winners’ career +/-:
Dirk Graham: -2
Rick Meagher: -45
Steve Kasper: -56

And maybe tell Brian Sutter (-63) and Ron Wilson (-42) that they were lousy defensive players too. Don’t forget Sammy Pahlsson: -35 = totally overrated defensive player.

Geez, I could go on…

Playing on shit teams = being a shit player.

You win the Internets, sir.

Clark: That lifetime plus-minus pretty much sums up his "all sizzle, no steak" career.

Wow, you have no clue what you’re talking about.

“This is rather a statistic that shows the overall value of a player on both the defense and offensive side of the situation.”

It still doesn’t adjust for QUALITY of competition, which is just as big of a problem.

Take Vanek’s +46 in 06-07. He did it against third and fourth line competition. When matched up against tougher opponents in 07-08 sans Briere/Drury, he had a HORRIBLE first half.