Why is +/- a bad stat?

[Filthy Dangles]

1) Because it's extrememly skewed and influenced by based on how good or bad your team is.

2) You get a minus for 5v6 goals when trying to tie up games late and that skews it too and vice versa, get a plus for 6v5 ENG. So, The stats can really get padded during garbage time.



But, context is key. Posters here use 5v5/EV GF% as if it's better than Plus-Minus despite being essentially the same thing.

 

[Cup or Bust]

Plus minus is a good stat only if used in the right context. For example looking at McDavid's +/- is extremely impressive this season when compared to the other top scorers in the league. The majority of the others play on better teams and teams that have much better goal differentials, gives a good indication of how dominant he is on the ice with less support. Last season on the Oilers for example Adam Larsson and Oscar Klefbom played as a defensive unit almost the entire season. Larsson finished +21 and Klefbom only +7. It certainly matched the eye test as Klefbom certainly had deficiencies in his defensive game last season, not just this season. Most of the success of that defensive unit defensively was due to Larsson playing better defensively compared to Klefbom, which was certainly recognizable when watching them. Even this season Larsson is +13 in a rough season on a team with -22 goal differential. People can say what they want about the Hall trade but Larsson absolutely makes the Oilers a better team. Leads the league in hits per game and is solid most nights.

Plus minus used in the scope of the team and the player usage is still a solid stat.

 

[Boom Boom Apathy]

Like any stat, it isn't bad, it is how it is used.

Exactly. Comparing +/- of a player from one team to a player on another team is completely useless. Players on bad teams are almost always going to have worse +/-.

Also, especially on bad teams, the guys with the best chance to score are going to be on late trying to push to catch up (and taking chances to do it) and even when they pull to goalie, so more often than not, the more "offensive" players are going to get the short end of the stick on +/-; again, this is more prevalent on bad teams. Other issues are when a team scores on the PP, but it counts as ES because a guy just stepped out of the box, or when a change is being made and a goal is scored (for or against) before the change is complete, etc... It also doesn't take into account quality of competition, quality of line-mates or zone starts.

It can be somewhat useful within a team when players are used similarly or you account for usage so it's not a "bad stat" per se, but as the poster above said, it's just how you use it.

 

[JetsWillFly4Ever]

It is bad for a multitude of reasons, many of which have already been covered.

1. It lacks predictive or descriptive value. Players can be +30 one year and -30 the next. It also only tells you who was on the ice when a goal was scored, not who was responsible.
2. It punishes players who play on the powerplay. It is only possible to get a minus while on the powerplay, not a plus.
3. It is skewed by empty net goals where you are pushing for a goal and don't really care if you get scored on.
4. It is largely based on how good your team is.
5. Goaltenders have a significant influence on it, not necessarily the players on the ice. If your goalie has a .930 when one player is on the ice but .890 when another is on the ice, the one that's on the ice when your goalies save percentage is really low is going to have a much worse plus/minus. There is little evidence that individual players influence goaltenders save percentage.

 

[Albi34]

It's bad because there are waaaay too many factors to consider on a team to team basis, let alone players within their own team. For example a forward line or d pair that always gets matched up against the opponents best offensive line is going to have a harder time gaining +- than the pair who defends the opponents 3rd line. the only thing i see a use for is looking at an entire teams +- as a whole rather than individual players.

 

[KingsFan7824]

Why is +/- considered such a bad stat?

As has been said, it's a stat that requires more context than most others.

and what do you consider to be the best stats?

All stats are a little wonky. They're all cumulative, but they may have nothing to do with the game you just played. If a 41 goal scorer scores 30 of them in the first 41 games, and 11 in the last 41, what is he? We say he scores .50 goals per game, but that's not necessarily true.

Game winning goals is another context free stat, like +/-. If you win a game 6-2, but were leading 6-0 with 10 minutes left in the game, which one of those 6 goals was the game winner? You could argue none of them were. Even if you win 3-2, but all the goals came in the 1st period, an argument can be made that there wasn't an actual game winning goal in that scenario.

Wins for a goalie fall into the same category. Most of the time, goalies don't win games by themselves. Like the game winning goal though, the stat has to go somewhere, so give it to the guy the plays the most minutes on a team.

I suppose that's why the whole advanced stats world has blown up. It's trying to give you more context to stats, but it hasn't figured it all out yet.

 

[Uncle Dru]

Year in and year I would love at least 5 of the 10 worst +/- players on my team. Obviously context is everything, but this day and age there are so many more sophisticated metrics that better measure/predict success

 

[Bleedred]

It fluctuates too much from year to year.

I remember Ovechkin got ripped that year that he was a -30-something while scoring 50 goals and a point per game. Meanwhile, that was one of only two seasons in his entire career and the first in 7 years in which he was a minus player. I remember people were like ''That's how much of a liability is Ovechkin is at Deefence!''.

It's a real dinosaur stat too. If you look at the leaders in plus/minus all time among defensemen, it's mostly guys who played either on dynasty teams or teams were damn near close to them. And most of them haven't been active in years. Jagr and Lidstrom (and Chelios in 2010, who played like a million years) are the only players to play a game this decade, who are in the top 25 all time in plus/minus. Then there's like less than 5 guys who are in the top 25 who have even played this century.

Only Bourque, Lidstrom, Stevens, MacInnis, Chelios and Jagr even played in this millennium.

 

[GellMann]

Jack Eichel: -18
Nic Deslauriers: +6

Better get Bergevin on the phone, Jason!!!

There is no usage context whatsoever, or context with respect to how good your team is, and very rarely is a random winger the sole reason why a goal gets allowed, or a random D playing an integral part in a goal scored, so many, many of the counts are "undeserved" when the actual play itself is looked at.

 

[Dr Jan Itor]

Every stat is fairly useless solely by itself.

 

[Vancouver Canucks]

When one person scores or is scored on, +/- of everyone on the ice(in the same line) is affected. +/- is a very inaccurate way of measuring player prowess.

 

[Flowah]

Hockey is the most random game of the major professional sports. That means, the stats you want to use to gauge a player are the ones where there are LOTS of events.

So shots, shot attempts. There are many dozens of those every game. There are literally games with only 1 goal total between the two teams and they're not exactly uncommon. A game with 6 total goals is above league average. That's not many events.

 

[jgatie]

+/- doesn't exist during the PP

Not totally true, a SH goal earns a plus.

 

[lomiller1]

+/- is just a bad implantation of goal differentials. Goal differential done properly is a much better stat but not the best stat. The main problem with a properly constructed goal differential stat is that goals are relatively uncommon, so even over a whole season luck doesn’t have enough time to even out.

+/- however is not properly constructed, use GF% instead. It covers the same things and does it properly.

It counts SH goals but not PP goals. Over the course of the season every PP will give up some goals and every PK will get some, so the guys on the PP will end up with a lower +/- and the guys who only get on the PK will end up with a higher +/-

It counts empty net goals. Over the course of a season every team will give up more empty net goals than they score 6v5. This means the players you put out 6v5 will get additional minus for being the most likely to score.

It does not normalize for time on ice. For example lets say a team has bad goaltending and lets in a lot more goals than they should. The players who get the most ice time will get more minuses that players who get less ice time.​

If you look at the totality, +/- severely punishes the best players on weak teams, and offensive players in particular. If you look at the worst +/- in the league the chances are that these are some of the best players, and if you look at the best you will frequently see below average players who get PK time on good teams.

As a rule of thumb:

If you want top understand a teams performance in games that have already been played use GF% and for players use Rel GF%.

If you want to understand how that same team will perform in upcoming games use CF% and for players use Rel CF%

 

[57special]

Two rock steady Dmen from the Wild, with their accompanying +/-'s during the past 4 years. Team made playoffs all 4 years.

Brodin; +21, -4, +5, +24

Suter; +7, +10, + 34, -1

I simply don't see any correlation between their +/- and their level of play, which has been more or less the same throughout that period.

 

[lomiller1]

+/- doesn't exist during the PP

You don't get a minus of you get scored on on a power play. The goal scorers don't get a plus either.

You get a "-" if your team gives up a either a SH goal or an empty net goal.

 

[izzy]

its not a bad stat

people use it in idiotic ways

its good when comparing players if you take into account their teams goal differential and the other players on their teams +/-

 

[lomiller1]

But isn't this partly the case with advanced stats as well?

It’s not a perfect solution, but most advanced stats use Rel numbers when comparing players from different teams. Rel numbers compare how the team does with the player on the ice vs when they are on the bench so they effectively measure whether they make the people one the ice with them better or worse.

 

[SotasicA]

+/- is better than Corsi. It's more objective and takes quality of shots inherently into consideration. A good measure of success in the game, in your role.

It's just not repeatable and not a good indicator of future performance. What it is, is a solid measure of past success. Performance and success are two very different things, so don't apply +/- without context. Jack Eichel performs well, but he does not have success, ie. other teams score on him more than he can score on them. A worse player can have success in a limited role on a different team, his line regularly beating the opposition. He can be a solid plus-player. It's not nothing. That player, in his surroundings, succeeds in the game of hockey, where the only real goal is to outscore the opposition.

It's a nice objective way to see if there are huge problems or great success happening. Kind of a team stat, but not really. If your top line has -30, that means they are getting outplayed. If your third line is +30, that means they are outplaying their opposition. Switch a couple of players around and suddenly your top line stops the bleeding and maybe suddenly your third line starts to allow a lot of goals. Who had success and who didn't? I think people who downplay +/- misunderstand it or just choose not to accept the info it gives. Your top scorer might have terrific scoring numbers, but if he's a minus-player, his line is constantly being outplayed. +/- isn't meant to indicate or predict whose fault it is they get minuses, just that they are.

 

[ulvvf]

Well it is more useful than corsi at least, since the game is about who score most goals not who shoot most shoots. But overall i think it is a pretty pointless stat. It is a team stat and really benifit medicore players.

But no matter what, i think it is more intresting how people are using it. Sometime they act like it is the most intresting stats above all, more than points or goals, and sometimes it is a pointles stats, all depending on what fits into their agenda to make one player look better or worse than he really is.

I think points and goals are the most intresting stats by far, nothing else comes close.

 

[lomiller1]

+/- is better than Corsi. It's more objective and takes quality of shots inherently into consideration. A good measure of success in the game, in your role.

The math has conclusively shown that any benefit from including shot quality information is massively outweighed by the noise from seeing eye shots, tips, shots that just make it in the net vs shots that hit the post, etc.

If you want to look at a team/player with luck included use GF%. If you want to look at how they would have performed without luck coming into play use CF%.

 

[ulvvf]

Not totally true, a SH goal earns a plus.

Yeah, and do not also the ones that have the powerplayer get a minus if they allow a goal? So the stats benifit medicore players. The ones that play in BP can not get a minus just plus, while it is the opposite for the ones that plays in PP.

 

[ulvvf]

It was a good stat years ago. But the fact is it almost always depends on how good your team is. there are some severe examples of people leading the league in Worst +/-, and then having the best after a trade because they went from a bad team to a good one. (Chara is the best example).

Why was it a good stat years ago but not now? What has changed?

 

[Alexander the Gr8]

Because when one guy goes -35 and is labeled a defensive liability, that same guy isn't praised for his defensive play when he went +45. People use this to push their agenda.

 

[lomiller1]

Well it is more useful than corsi at least, since the game is about who score most goals not who shoot most shoots. But overall i think it is a pretty pointless stat. It is a team stat and really benifit medicore players.

But no matter what, i think it is more intresting how people are using it. Sometime they act like it is the most intresting stats above all, more than points or goals, and sometimes it is a pointles stats, all depending on what fits into their agenda to make one player look better or worse than he really is.

I think points and goals are the most intresting stats by far, nothing else comes close.

Whether you care about past goals or future goals is key. GF% most accurately describes what goals occurred, CF% outperforms it for predicting what goals will be scored. +/- is not GF%, it’s just a broken useless statistic. Don’t use it period.