+/- is a poop stat and there's video proof too!

[fanaticV3.0]

http://www.broadstreethockey.com/2015/1/19/7854163/nhl-plus-minus-stat-awful-flyers-sean-couturier-breakdown

 

Solid article on why this terrible stat has no value.

[TropicalFruitGirl26]

I wouldn't call +/- TOTALLY worthless, but i would have to agree that not a ton of weight should be put into it.

 

Personally, I use it along with a composite of other stats when trying to evaluate the effectiveness of a given player's defensive capabilities.

Furthermore, I use it on a case by case basis with certain players (because face it, some players simply play more and against tougher competition than others), and put a bit more importance (not much more mind you) if the player is an actual defenseman or forward as I think the position skews the value of it.

 

So, when I look at, say a given D-man, and want to get a broad picture of what he could possibly bring year in, year out on the ice, I look at things like his blocked shots ALONG with, hits, ALONG with giveaways/takeaways, along with average TOI, and after looking at those, see what teams he played on, what that team has for overall defense and goaltending, THEN maybe have a look see at what his +/- is while playing for that team, and given his other numbers as a whole.

 

For forwards, I almost NEVER look at +/-, as a forwards primary job is to score goals, or at least, muck up the works (for checking forwards) so that the other team has trouble scoring themselves.

I get that its part of their jobs to back check and help on defense as well, but again, that is what you have DEFENSEMEN and a goalie for, as stopping scoring and scoring chances is their primary job.

 

And with many top flight forwards going more all in on offense with maybe a smidge of defense, AND if they play a ton of minutes, then chances of having scoring chances against are much greater....especially if the given defense or goalies don't prevent the goals, which is something that is mostly out of the forwards' control.

 

I've said for a long time that some ppl put way too much stock in +/-.

It certainly can be used for certain situations, particularly if a given players actions lead directly to goals (and it happens pretty consistently with that player on top of that)....or the opposite, where a given players actions are directly respsonsible for scuffed shots or missed shots...THEN in those particular instances, I would say a +/- rating could be more telling.

 

But for those that would hold certain players in high regards based solely on the actual +/- number with little to no regard to other stats or situations that lead to that number, I say, "get a grip"

[Podein25]

Tell us something that we didn't already know.

 

The stat is just a proxy. No more, no less. It has never been anything else. As it turns out, an imperfect one. Anyone who has ever played the game knows this since, odds are, they've experienced the blind luck of it. 

 

And yet it works just fine as a proxy: Toews is +15, Kessel is -14. Funny that.

[JackStraw]

Tell us something that we didn't already know.

 

The stat is just a proxy. No more, no less. It has never been anything else. As it turns out, an imperfect one. Anyone who has ever played the game knows this since, odds are, they've experienced the blind luck of it. 

 

And yet it works just fine as a proxy: Toews is +15, Kessel is -14. Funny that.

 

Correctly correct. It's fine if you don't expect it to tell you things that it can't. Larger sample size also is useful, I don't see to many crappy players among the all time leaders:

 

http://www.hockey-reference.com/leaders/plus_minus_career.html

 

Meanwhile, everybody is going weak in the knees over Corsi, which is just plus-minus for shots instead of goals. Go figure.

[radoran]

But for those that would hold certain players in high regards based solely on the actual +/- number with little to no regard to other stats or situations that lead to that number, I say, "get a grip"

 

QFT

 

It is a tool and what matter is how you use that tool. Taken in context as an overall evaluation of a player, the stat can certainly have some value.

 

Agreed wholeheartedly that for forwards it has very low value (Lindros and LeClair didn't lead the league in +/- because of their "defensive acumen"), but even the very article posted shows evidence of why it can be of value for a defenseman.

 

Video one: Grabner "flies by" Coliavcovo - does Coliacovo "deserve" a - for that?

Video two: Kulemin "lost in front" by #22 standing right there. - does Schenn "deserve" a - for "losing" Kulemin?

Video four: puck off MDZ's skate standing in front - does MDZ "deserve" a - for that?

 

I'd say that, yes, these defensemen earned their "minus".

 

Quite frankly, I do think that counting +/- away on empty netters is also ludicrous.

 

Again, it's how you use the stat. If one wants it to be the be all/end all of a player, one will be Very Disappointed Very Quickly.

 

Applying the stat to one game is also a fool's errand.

 

For example, Couturier even with the -5 in that game is even for the season. He's also had 13 games where he was a + player and 13 games where he was a - player. So, he's even for the season despite the -5.

 

That tells you a whole lot more about the player - that he could still be "even" despite the -5 - than simply looking at the -5 and saying "he was terrible" does.

 

FWIW, the media guys are also complete idiots. But we knew that.

[Podein25]

Correctly correct. It's fine if you don't expect it to tell you things that it can't. Larger sample size also is useful, I don't see to many crappy players among the all time leaders:

 

http://www.hockey-re...nus_career.html

 

Meanwhile, everybody is going weak in the knees over Corsi, which is just plus-minus for shots instead of goals. Go figure

 

 

Exactly exact. It's impossible, over the long term, to be a lousy player (by which I mean an incomplete player, a one-dimensional player, an offence-only player, a lazy player etc) and have a positive plus. The puck ends up in the net against guys like that. Eventually. And again. It just does.

[fanaticV3.0]

I wouldn't call +/- TOTALLY worthless, but i would have to agree that not a ton of weight should be put into it.

 

Personally, I use it along with a composite of other stats when trying to evaluate the effectiveness of a given player's defensive capabilities.

Furthermore, I use it on a case by case basis with certain players (because face it, some players simply play more and against tougher competition than others), and put a bit more importance (not much more mind you) if the player is an actual defenseman or forward as I think the position skews the value of it.

 

So, when I look at, say a given D-man, and want to get a broad picture of what he could possibly bring year in, year out on the ice, I look at things like his blocked shots ALONG with, hits, ALONG with giveaways/takeaways, along with average TOI, and after looking at those, see what teams he played on, what that team has for overall defense and goaltending, THEN maybe have a look see at what his +/- is while playing for that team, and given his other numbers as a whole.

 

For forwards, I almost NEVER look at +/-, as a forwards primary job is to score goals, or at least, muck up the works (for checking forwards) so that the other team has trouble scoring themselves.

I get that its part of their jobs to back check and help on defense as well, but again, that is what you have DEFENSEMEN and a goalie for, as stopping scoring and scoring chances is their primary job.

 

And with many top flight forwards going more all in on offense with maybe a smidge of defense, AND if they play a ton of minutes, then chances of having scoring chances against are much greater....especially if the given defense or goalies don't prevent the goals, which is something that is mostly out of the forwards' control.

 

I've said for a long time that some ppl put way too much stock in +/-.

It certainly can be used for certain situations, particularly if a given players actions lead directly to goals (and it happens pretty consistently with that player on top of that)....or the opposite, where a given players actions are directly respsonsible for scuffed shots or missed shots...THEN in those particular instances, I would say a +/- rating could be more telling.

 

But for those that would hold certain players in high regards based solely on the actual +/- number with little to no regard to other stats or situations that lead to that number, I say, "get a grip"

 

I would and I'm not speaking in hyperbole either.

 

The rest of your post, which is laid out nicely, is why. You were very clear about when you apply it and find it to have value. That's not how stats (should) work. Stats are facts. What does this particular fact record? That a player was on the ice when his team scored or gave up a goal. This information has no use, because as demonstrated  in the article (and 1000s of times throughout a given season) there are always players on the ice who have nothing to do with the play in question. But this stat gives them credit or punishes them for simply being there. It's like guilt by association. If you have to dance and juggle to come up with an explanation as to what players it applies to and when it has value, it has no value. 

[fanaticV3.0]

Tell us something that we didn't already know.

 

The stat is just a proxy. No more, no less. It has never been anything else. As it turns out, an imperfect one. Anyone who has ever played the game Anyone who has functional eyes, follows the sport closely, or even just has the tiniest bit of common sense knows this since, odds are, they've experienced the blind luck of it playing the game is irrelevant to one's ability to speak about it.

 

And yet it works just fine as a proxy: Toews is +15, Kessel is -14. Funny that.

 

I fixed your statement for you. It had a logical fallacy in it. You're welcome.

[fanaticV3.0]

Correctly correct. It's fine if you don't expect it to tell you things that it can't. Larger sample size also is useful, I don't see to many crappy players among the all time leaders:

 

http://www.hockey-reference.com/leaders/plus_minus_career.html

 

Meanwhile, everybody is going weak in the knees over Corsi, which is just plus-minus for shots instead of goals. Go figure.

 

I think you meant to type "dumb people" there. Corsi might be the only stat that is a bigger load of sh-t than the plus/minus.

[radoran]

By this "analysis" the "assist" numbers are worthless because sometimes an assist is awarded to guys who really had nothing to do with the goal itself.

 

Or, SV% is worthless because some teams give up more shots than others.

 

Or, "shot blocks" are worthless because it doesn't take into account that guys allow players to shoot more.

 

Heck, "goals" could be "worthless" because some are scored on empty nets. Or off a random skate/stick deflection. Or off a defenseman's butt. Or a guy gets "credited" with a goal he never even touched.

 

Stats are what one makes of them. No stat tells "the whole story" about a player or a team.

 

Never has. Never will.

 

"Hockey people" apply +/- as part of player evaluation because over the course of decades of experience it has been shown to have value.

 

As pods said:

 

Exactly exact. It's impossible, over the long term, to be a lousy player (by which I mean an incomplete player, a one-dimensional player, an offence-only player, a lazy player etc) and have a positive plus. The puck ends up in the net against guys like that. Eventually. And again. It just does.

 

No matter what a Couturier fanboy on the Broad Street Report thinks.

[Zz_]

Only reason you should look at +/- is to see who is playing solid offense when/defense on the forward position. Otherwise, most commonly know are the most unwanted stat in the history of hockey. 

[WordsOfWisdom]

+/- is the most misunderstood stat in hockey.

 

People think it measures defence and they continue to use it as a defensive stat.

 

+/- is meant to be an OVERALL MEASURE OF A PLAYER's WORTH TO HIS TEAM. It can only really be viewed within the context of a single team, and even then, it measures the performance of the LINE more than it measures the performance of the INDIVIDUAL.

 

I created the Defensive Errors (DE) statistic to address all of the concerns that +/- does not handle.

 

 

Comparison of DE to +/-:

 

Offence:

 

+/- (YES)

DE (NO)

 

Defence:

 

+/- (YES)

DE (YES)

 

Line specific:

 

+/- (YES)

DE (NO)

 

Affected by team performance:

 

+/- (YES)

DE (NO)

 

Summary:

 

- Player that leads the league in +/- is a great offensive player on a great overall team.

- Player that leads the league in MPBE (Minutes Played Between Defensive Errors) is the best defensive player in the NHL.

[Polaris922]

But who is deciding what constitutes a defensive error? It becomes purely subjective which instantly builds in error.

[Podein25]

+/- is the most misunderstood stat in hockey.

People think it measures defence and they continue to use it as a defensive stat.

+/- is meant to be an OVERALL MEASURE OF A PLAYER's WORTH TO HIS TEAM. It can only really be viewed within the context of a single team, and even then, it measures the performance of the LINE more than it measures the performance of the INDIVIDUAL.

Yes. It's a good point about how a *line* performs - as a line. I've never seen the stat as a measure of defense per se, buy it always served as a decent proxy for whether forwards gave a crap about their own end or the defensive side of the puck.

[WordsOfWisdom]

Yes. It's a good point about how a *line* performs - as a line. I've never seen the stat as a measure of defense per se, buy it always served as a decent proxy for whether forwards gave a crap about their own end or the defensive side of the puck.

 

That's basically the idea. If a player can get ~80 points in this league and still be -20, it speaks volumes about the lack of defensive ability of that line.

 

A minus player is a player that hurts his team more than he helps them when he's on the ice. (But again, it's line specific and influenced by other factors like quality of goaltending, power plays, etc.)

 

[JagerMeister]

I love when people use +/- to judge how good someone is defensively. I guess Jagr(+290) and Gretzky(500) are the best two way players ever.

[Podein25]

No offense, but I don't think you appreciate how many hockey people have traditionally used that stat as a rough gauge *precisely* for whether a player had a good two way game or not.

[JagerMeister]

No offense, but I don't think you appreciate how many hockey people have traditionally used that stat as a rough gauge *precisely* for whether a player had a good two way game or not.

Well, do you really think Jagr and gretzky where good two way players? Ron Francis is a career -10 yet he was a good two way player, he even won a selke. Ovechkin is a career plus 59, even though everyone was complaining how bad he was defensively because of his -35 lol. Now he is +12. you dont see those people complaining that he is bad defensively anymore.

[Podein25]

Well, do you really think Jagr and gretzky where good two way players? Ron Francis is a career -10 yet he was a good two way player, he even won a selke. Ovechkin is a career plus 59, even though everyone was complaining how bad he was defensively because of his -35 lol. Now he is +12. you dont see those people complaining that he is bad defensively anymore.

A fair point, but I never said it was a prefect stat. It is a far way from that. I still think it is a good rough measure for how much offensive players care about the defensive side of the puck.

Similarly, a purely defensive player who never gets on the score sheet, would have a low plus (or a minus), and that too would be a decent indicator that he has a poor two way game.

[JagerMeister]

A fair point, but I never said it was a prefect stat. It is a far way from that. I still think it is a good rough measure for how much offensive players care about the defensive side of the puck.

Similarly, a purely defensive player who never gets on the score sheet, would have a low plus (or a minus), and that too would be a decent indicator that he has a poor two way game.

You cant do that the other way around though, because some purely offensive players have high +/- Like Jagr, gretzky and a few more players im too lazy to name. Im just saying that people shouldnt use the stat to show how bad someone was defensively.

[WordsOfWisdom]

But who is deciding what constitutes a defensive error? It becomes purely subjective which instantly builds in error.

 

Well, therein lies the rub.    

 

Since defensive errors are of my own creation and not an official hockey stat (nobody outside of this forum has ever heard of it), it means that I get to define what it is and how to assign it. (Which I did in the original post of that thread.)

 

Now, could better hockey minds with better TV/video camera angles do a better job of achieving what I've set out to achieve so far this season? Probably. I think the point though, is that it is possible to assign the errors in a consistent manner, just like goals and assists are assigned to players.

 

If you look at the NHL stats used today, here they are from most ironclad to most subjective:

 

1. Wins & Losses (ironclad)
2. Goals (ironclad -- video reviewable)
3. Assists (should be ironclad)
4. +/- (ironclad, provided player tracking is accurate)
5. ATOI / TTOI / MIN (ironclad, provided player tracking is accurate)
6. Shots On Goal (scorekeeper's discretion)
7. Penalties In Minutes (referee's discretion, no two refs are the same)
8. Giveaways / Takeaways / Blocked Shots (scorekeeper's discretion)

 

I think DE could be as ironclad as #4 on that list, and certainly no lower than #6.

 

It would take training for people to learn how to score it correctly. But once that initial learning curve is done, it would become as much a part of the game as goals and assists. I would love it if the in-game score ticker showed something like this when a goal is scored:

 

Example:

 

1: 13:28: PIT - GOAL: Crosby (Malkin), TOR - ERROR: Franson, Kadri.