Early examples of why +/- is the worst stat in hockey and fantasy hockey

[rataylor22]

http://youtu.be/1abp697-w4c

http://youtu.be/2hXMx_aBQo4

If you'll kindly follow those links, you'll see 9 of 10 players did not deserve a plus on either play (only O' Reilly did). Conversely 8 of 9 players did not deserve a minus on either play (only Beauchemin did).

It is a dumb and awful stat.

[deantime419]

Yes, we know the arguments about how plus/minus and PIMs are bad stats in fantasy hockey due to their uselessness and negative impact in real hockey games BUT at the same time they generally seem to follow enough of a trend that from a pure fantasy hockey standpoint, they somewhat make sense to use as scoring categories.

On a side note though, being that people are incorporating stats such as hits instead of PIMs, has anyone ever been in a league that used takeaways/giveaways as a stat instead of plus/minus?

[LakeSoup]

I don't think Takeaways/giveaways is an option on Yahoo custom. But that's a brilliant idea.

[Mr. Guru]

you win as a team you lose as a team

[Pengwin7]

I seem to ALWAYS be outvoted/out-opinioned... but... PLUS/MINUS is my favourite stats of all hockey stats.
I truly believe that, over the course of somebody's career (not week, not month... sometimes year), a player's true NHL value can be viewed statistically through one's +/- relative to their teammates.
(RELATIVE TO THEIR TEAMMATES: BOLD, UNDERLINED... PEOPLE CONTINUE TO NOT BE ABLE TO TAKE IN THAT PART!!!)

This view, for me, is two-fold:
1) Almost 30 years of actual hockey games I've played.
The hardest working two-way players on my teams ALWAYS had the best +/- over the course of a full season.

2) Just look at the top +/- guys in the history of the NHL:
1. Larry Robinson* 730
2. Bobby Orr* 597
3. Raymond Bourque* 528
4. Wayne Gretzky* 518
5. Bobby Clarke* 506
6. Denis Potvin* 460
Serge Savard* 460
8. Guy Lafleur* 453
9. Bryan Trottier* 452
10. Nicklas Lidstrom 450

Can somebody really look at that list and tell me that these aren't incredible players?
And it's not just their rank in the NHL... it's that this rank also reflects how incredible they were relative to their own teammates (who may not be anywhere close to their rank).

I'm sure I'll stir a debate... as I usually do with plus/minus... but I find that people that bash the +/- either:
i) Didn't play much actual hockey growing up, and/or
ii) Don't understand the value in back-checking, and/or
iii) Don't understand how to focus on +/- relative to one's teammates.


I realize this is a fantasy hockey website... but there is real value in +/-... coaches know it, GMs know it, scouts know it.
Only the casual hockey fans doesn't see the value in plus/minus.
Sorry to hurt some feelings out there... but I strongly disagree with anybody dissing +/- as a stat.
Tell that to Patrice Bergeron, Pavel Datsyuk, and Jon Toews.

[Axeman33]

The stat I always found to ironic in fantasy hockey is PIMs.

If you think about it, in real hockey when you get a penalty you put your team at a disadvantage. Yet in fantasy hockey, guys who get PIMs on your roster offer your team positive points.

I realize it makes guys who normally wouldn't have fantasy value valuable but the irony behind that stat catagory always had me wondering...

[LawMan]

I seem to ALWAYS be outvoted/out-opinioned... but... PLUS/MINUS is my favourite stats of all hockey stats.
I truly believe that, over the course of somebody's career (not week, not month... sometimes year), a player's true NHL value can be viewed statistically through one's +/- relative to their teammates.
(RELATIVE TO THEIR TEAMMATES: BOLD, UNDERLINED... PEOPLE CONTINUE TO NOT BE ABLE TO TAKE IN THAT PART!!!)

This view, for me, is two-fold:
1) Almost 30 years of actual hockey games I've played.
The hardest working two-way players on my teams ALWAYS had the best +/- over the course of a full season.

2) Just look at the top +/- guys in the history of the NHL:
1. Larry Robinson* 730
2. Bobby Orr* 597
3. Raymond Bourque* 528
4. Wayne Gretzky* 518
5. Bobby Clarke* 506
6. Denis Potvin* 460
Serge Savard* 460
8. Guy Lafleur* 453
9. Bryan Trottier* 452
10. Nicklas Lidstrom 450

Can somebody really look at that list and tell me that these aren't incredible players?
And it's not just their rank in the NHL... it's that this rank also reflects how incredible they were relative to their own teammates (who may not be anywhere close to their rank).

I'm sure I'll stir a debate... as I usually do with plus/minus... but I find that people that bash the +/- either:
i) Didn't play much actual hockey growing up, and/or
ii) Don't understand the value in back-checking, and/or
iii) Don't understand how to focus on +/- relative to one's teammates.


I realize this is a fantasy hockey website... but there is real value in +/-... coaches know it, GMs know it, scouts know it.
Only the casual hockey fans doesn't see the value in plus/minus.
Sorry to hurt some feelings out there... but I strongly disagree with anybody dissing +/- as a stat.
Tell that to Patrice Bergeron, Pavel Datsyuk, and Jon Toews.

Haha. As soon as I saw the title I knew this post was coming from Pengwin, I gotta respect your love for +/- in the face of all the hatred. Also, you do make some solid points.

Now to play devil’s advocate to both sides:
1. The top-10 guys in +/- are amazing hockey players but correlation does not equal causation. For a comparison I’d argue GWG is the worst hockey stat out there, fantasy or otherwise, because it simply does not measure anything other than luck. The top-10 players in NHL GWG all-time are:

1 Jaromir Jagr
2 Phil Esposito
3 Brett Hull
4 Brendan Shanahan
5 Teemu Selanne
6 Guy Lafleur
7 Mats Sundin
8 Steve Yzerman
9 Sergei Fedorov
10 Joe Nieuwendyk

(note: there is a boost to those who played during the regular season OT era). These are also all great hockey players. But they are great hockey players who scored a lot of game winning goals, they aren’t great hockey players because they are top-10 in GWG rather they just happen to be both great hockey player and top-10. Those top-10 +/- guys are near the top in a lot of statistical categories, some even more obscure than +/-.

2. I agree with you that +/- relative to teammates is a good indication of a player’s strength as it isolates the player’s contributions, but that stat is not used in fantasy hockey and is rarely talked about. The biggest problem with +/- is that it’s a team driven stat used to argue an individual player’s ability, much like goalie wins it gives us a better understanding of the team’s play then the individuals.

3. +/- has been superseded by other advanced stats which better reflect a player’s value. This is not a fantasy argument but an NHL argument. Hockey stats are still well behind baseball and others but with new advanced stats we get a better picture than what +/- gives us.

I’m actually middle of the road on +/-. I think it has some value over the course of a season, and in fantasy I have no problem using it as is reasonably predictable. However, just like any single stat it has its limitations, namely, it punishes guys on bad teams (or with bad goalies) and flatters guys on good teams. I consider +/- to be the goalie wins of the skater category.

Finally, in support of +/-,or any stat you can find a few clips to show the stat being awarded wher it wasn’t deserved by the player that doesn’t make the entire stat worthless. By the OP’s argument I can confidently say that neither McDonagh, nor Boyle, deserved a goal or assist respectively on the Quick goal yet those points were still awarded. Despite this one bad representation of the stat I’m sure we can all agree that goals and assists are very important stats to keep.

[27Blue]

+/- is a useless stat...

even when comparing players on the same team it is useless without context. you would actually have to go and watch a team play for an extended period to understand why the players have the +/- stats they have. linemates, situations used in, quality of opponents all factor in. there are too many variables to get a picture by just looking at the stat, or even by just looking at the stat in relation to teammates.

i guess the same could be said about most stats, though... even points.

[robmyatt]

with any stat, sample size is the most important consideration.
Over a course of a career it tells a lot.
Over a course of a season it tells a little.
Over the course of a week or month it means nothing.

It is affected by too many other factors. That's why I prefer a stat like corsi or fenwick because it enables you to measure "puck posession" (which correlates very strongly with a winning team) against players on your own team and even line.

[doulos]

Brian Burke called it a "horse shit" stat. Thus it has been decided

[Pengwin7]

The biggest problem with +/- is that it’s a team driven stat used to argue an individual player’s ability, much like goalie wins it gives us a better understanding of the team’s play then the individuals.

3. +/- has been superseded by other advanced stats which better reflect a player’s value. This is not a fantasy argument but an NHL argument. Hockey stats are still well behind baseball and others but with new advanced stats we get a better picture than what +/- gives us.

You understand it well.

I love advanced stats, but even like +/- almost every stat is team related. For example, if Parise/Koivu play together on every shift of every game, there is nothing to separate one's contribution from the other. In fact, part of my intense hatred for Ilya Kovalchuk (or Michael Grabner) is that he consistently need a forward linemate that would play a defensive game simply to prevent the line from getting scored upon so often.

It's funny that people bring up +/- so often... but rarely bring up GAA for goalies.
GAA is basically just a goalie's SV% factored in with the number of shots a team gives up (which is a factor of "team" defense).

Anyways, I don't really have a counter-point to make to you.
You understand the positives/negatives with this stat quite well.

+/- is a useless stat...

even when comparing players on the same team it is useless without context. you would actually have to go and watch a team play for an extended period to understand why the players have the +/- stats they have. linemates, situations used in, quality of opponents all factor in. there are too many variables to get a picture by just looking at the stat, or even by just looking at the stat in relation to teammates.

i guess the same could be said about most stats, though... even points.

It can be a useless stat without context, yes.
But the context is that we ARE hockey fans and we DO watch the games.
Most of us KNOW why Jeff Schultz was +50 one year.
Those of us that do NOT know why Jeff Schultz was +50 one year... well, those are the people that have no context and simply go pointing to +/- as a useless stat.

Depending upon somebody's upbringing with hockey, they learn to value/devalue certain things.
Sometimes I like to ask people what they feel are the five most important occupations.
Then I ask them how well they would survive in a raw world with five people of those occupations.
It's amazing what knowledge/function some people value and some people do not.

with any stat, sample size is the most important consideration.
Over a course of a career it tells a lot.
Over a course of a season it tells a little.
Over the course of a week or month it means nothing.

It is affected by too many other factors. That's why I prefer a stat like corsi or fenwick because it enables you to measure "puck posession" (which correlates very strongly with a winning team) against players on your own team and even line.

Yes, agreed.

But we are talking about statistics for fantasy hockey... and there is only one skater category that can best factor in a player's defensive play... and that is +/-. The casual hockey fan has no idea how many bad passes Patrice Bergeron may force by simply getting in opposing passing lanes. How could somebody quantify that value? +/- is the closest fantasy statistic we have towards that real value.

[grendel]

Brian Burke called it a "horse shit" stat. Thus it has been decided

and Brian Burke is unquestionably an expert when it comes to shit

[doulos]

But we are talking about statistics for fantasy hockey... and there is only one skater category that can best factor in a player's defensive play... and that is +/-. The casual hockey fan has no idea how many bad passes Patrice Bergeron may force by simply getting in opposing passing lanes. How could somebody quantify that value? +/- is the closest fantasy statistic we have towards that real value.

I think the point many are making isn't that it's the best stat we have for tracking defensive awareness (in a fantasy hockey setting), but that it's such a useless stat that it shouldn't be used at all in a fantasy hockey setting.

When 30-40% of the +/- scoring is given to the wrong player (this was the number found on a 5 year study done on the Oilers) then it really puts into question the usefulness of the stat even in comparing between players on the same team.

However, comparing between players on the same team is not how the stat is used in fantasy hockey. It's used as a comparison between players in the whole league, so even if it's being touted as a 'bad stat but the best we have', it's not even being used properly.

The only thing +/- shows for fantasy hockey purposes, is which players play on good vs bad teams. If that's the reason to use the stat then I can see it (though that seems silly to me personally), but it's wrong to say that it represents defensive abilities - in a fantasy hockey setting - when it actually does no such thing.

Saying that someone has to have a long background in playing hockey to understand plus/minus is totally silly. People can watch hockey and understand the game without having athletic ability.

Taylor Hall is lagging the play and sneaks onto the bench. Boyd Gordon jumps over the boards, and seconds later the team is scored on - mainly because Taylor Hall was out of position and left the ice. Yet, Boyd Gordon gets the big minus for this. Notice I don't need to be great on my skates to understand and see what happened there.

To the point that someone who doesn't understand +/- doesn't appreciate the value of backchecking, I'd just point out the example I just showed. Quite the contrary, the full understanding of backchecking is exactly why +/- is a completely flawed stat.

Finally you said:

"I realize this is a fantasy hockey website... but there is real value in +/-... coaches know it, GMs know it, scouts know it."

So, now we're going to move away from the world of fantasy hockey (in which I see no place for +/- as anything other than establishing how good the team is that a player plays for) and into the real world.

I wasn't making it up that Brian Burke said it was a horse-shit stat - it's true.

So two things.

1) Do coaches/GMs/scouts REALLY value +/- to such a degree like you seem to insinuate? I'm not convinced this is true.

2) Even if they DO, that doesn't establish anything, except that they value it. There are lots of things in many fields that are so totally bogus but people cling to because it's considered 'established wisdom/knowledge'

[27Blue]

i guess that settles it... if you don't play bush league hockey, then you don't understand the game. from now on i am only taking advice from 12 year old kids who play minor hockey and beer leaguers who obviously have an inside edge.

thanks.

[Pengwin7]

So two things.

1) Do coaches/GMs/scouts REALLY value +/- to such a degree like you seem to insinuate? I'm not convinced this is true.

2) Even if they DO, that doesn't establish anything, except that they value it. There are lots of things in many fields that are so totally bogus but people cling to because it's considered 'established wisdom/knowledge'

Do you realize what you wrote here?

1) Prove to me that intelligent hockey minds believe +/- is valuable.
2) Even if you prove it... I have written this second clause that shall justify my previous & future disregard for value of this statistic.


...c'mon.