Most useless or inaccurate stat in hockey?

[VictoriaJetsFan]

Need to defend plus/minus here...

Given the right context, it can be highly valuable. When one guys pus/minus is greatly different from the rest of his team, it makes a statement, it is worth something.

 

[Jumptheshark]

for me, Points or goals per game.

My fellow oiler fans keep saying Ales Hemsky is nearly a pt per game player. The problem with that statement is that Hemsky gets injured a lot ( we all know that and we need to admit it)

I will use the 2010/2011 season as an example

Based upon how we do pts per game Hemsky had a .89ppg average that year. One problem. He played in on 47 games out of the 82. If we use the 82 game schedule, that takes his PP down .51.

or if you look at his career he is a life time .76ppg player--but here is the problem with that stat. He has missed between 300 to 350 games in his career and that takes his PPG stat down to about .5 when you factor his actual point out out and use the number of games he could have played in

The PPG stat hides how many games he actual played in vs how many he could have

 

[octopi]

Plus/Minus.

That stat is more about a players team than the player himself.

and doesn't take PP or PK into account, so there's that too.

I think SOG while not useless is very deceptive. A team can have 50 shots, but they can be of mostly very poor quality.

 

[that is unfortunate]

GAA for goalies.

 

[Evincar]

Wins and GAA for goalies.

Individual Corsi.

and doesn't take PP or PK into account, so there's that too.

I think SOG while not useless is very deceptive. A team can have 50 shots, but they can be of mostly very poor quality.

But if a team has 50 shots it should mean that they are controlling the play/are in the other team's zone a lot.

 

[cujoflutie]

+/- isn't a useless stat I just find it gets misinterpreted quite a bit

a lot of people assume a players defensive ability is determined by +/- which is untrue. Common sense says half that stat is based on team offense but people often fail to take into account the situation the player is in; for instance Mike Green has had some solid +/- seasons despite being terrible defensively mainly due to his team's offense. Likewise the best defenders on bad teams will usually have terrible +/-'s because they are logging heavy minutes against the other team's top units. It also doesn't take into account ice time.


I wish they'd come up with a GAA stat for skaters; not hard to do, calculate goals against at even strength/PP and divide by icetime. Or divide icetime by goals against to have a minutes per goal against stat.

 

[Blue Blooded]

+/- isn't a useless stat I just find it gets misinterpreted quite a bit

a lot of people assume a players defensive ability is determined by +/- which is untrue. Common sense says half that stat is based on team offense but people often fail to take into account the situation the player is in; for instance Mike Green has had some solid +/- seasons despite being terrible defensively mainly due to his team's offense. Likewise the best defenders on bad teams will usually have terrible +/-'s because they are logging heavy minutes against the other team's top units. It also doesn't take into account ice time.


I wish they'd come up with a GAA stat for skaters; not hard to do, calculate goals against at even strength/PP and divide by icetime. Or divide icetime by goals against to have a minutes per goal against stat.

http://stats.hockeyanalysis.com/rat...=500&teamid=0&type=goals&sort=A20&sortdir=ASC

 

[Boom Boom Bear]

Assists are the most overly weighted, inaccurate abstraction of a statistic. The mere existence of the assist defies basic logic. What is being counted: shots or passes? Why is the assist worth an individual point equal to the value of a goal? Why are there two assists awarded? Why is setting up a goal worth twice the value of the goal itself?

When one reads a player's assist totals, what does it actually say about his game? Breaking down between first & second assists still tells us nothing about whether someone is shoot first or pass first, & it tells us virtually nothing about a person's shooting or passing ability other than the fact that they touched the puck shortly before their teammate scored. Period.

Get rid of assists & points & replace them with more accurate stats, like maybe

* completed passes
* shot assists, or shots leading to goals [ie deflected or tipped in]

 

[do0glas]

Assists are the most overly weighted, inaccurate abstraction of a statistic. The mere existence of the assist defies basic logic. What is being counted: shots or passes? Why is the assist worth an individual point equal to the value of a goal? Why are there two assists awarded? Why is setting up a goal worth twice the value of the goal itself?

When one reads a player's assist totals, what does it actually say about his game? Breaking down between first & second assists still tells us nothing about whether someone is shoot first or pass first, & it tells us virtually nothing about a person's shooting or passing ability other than the fact that they touched the puck shortly before their teammate scored. Period.

Get rid of assists & points & replace them with more accurate stats, like maybe

* completed passes
* shot assists, or shots leading to goals [ie deflected or tipped in]

i agree about 1st assist/2nd assist.

however, to some extent. i can look at A/60 along with S/60 and get an idea about the player.

for instance:

Tyler Kennedy. 11/12-4th on his team in A/60. his shot/60? 13.8!!

joe thornton 1st on his team in A/60 and his S/60 is like 5.

its a general idea, nothing too precise. but its what i use.

 

[habsfanatics*]

Corsi. I'm still waiting for Gomez to
Rebound.

 

[MrClutch86]

+/- defiantly.

 

[DyerMaker66*]

Assists are the most overly weighted, inaccurate abstraction of a statistic. The mere existence of the assist defies basic logic. What is being counted: shots or passes? Why is the assist worth an individual point equal to the value of a goal? Why are there two assists awarded? Why is setting up a goal worth twice the value of the goal itself?

Because said goal does not exist without said assist.

 

[Brainiac]

Most inaccurate: anything that requires interpretation, like giveaways, takeaways.

Most useless: that one is almost impossible to answer. Most metrics, if interpreted properly, can be useful. If I had one to pinpoint, I'd say game winning goals. Very accurately measured, but extremely flukey.

And a big LOL to all of you saying +/- is useless. +/- is way overused by many reporters, but it's far from useless.

 

[Pip]

Wins for goalies

 

[BadgerBruce]

Any statistic related to the Shootout always makes me think of 8 year olds playing "Showdown" in their parents' basement. Until shootouts are used in Game 7 of SC finals, stats associated with the gimmick tell me nothing of value about teams or individual players.

 

[Blue Blooded]

Any statistic related to the Shootout always makes me think of 8 year olds playing "Showdown" in their parents' basement. Until shootouts are used in Game 7 of SC finals, stats associated with the gimmick tell me nothing of value about teams or individual players.

Had the Kings gone 0-15 instead of 6-9 in the shootout they would have missed the playoffs in 2011-12.

Had the Flyers lost the shootout in the final game of the reg. season in 2009-10 they would have missed the playoffs and never made the finals that year.

 

[Cunneen]

Had the Kings gone 0-15 instead of 6-9 in the shootout they would have missed the playoffs in 2011-12.

Had the Flyers lost the shootout in the final game of the reg. season in 2009-10 they would have missed the playoffs and never made the finals that year.

Shootouts are basically a crapshoot, and differences in teams/players are usually the result of random chance.

http://www.arcticicehockey.com/2010/12/29/1901842/nhl-shootout-is-a-crapshoot

http://www.broadstreethockey.com/2012/2/13/2789709/yes-it-is

 

[Mrb1p]

Any goaltending stats. Keeping goals away from your net is a team thing.

 

[GuineaPig]

Any goaltending stats. Keeping goals away from your net is a team thing.

But there isn't much evidence to suggest that teams vary that much in this respect at even strength.

 

[TheDevilMadeMe]

Shootouts are basically a crapshoot, and differences in teams/players are usually the result of random chance.

http://www.arcticicehockey.com/2010/12/29/1901842/nhl-shootout-is-a-crapshoot

http://www.broadstreethockey.com/2012/2/13/2789709/yes-it-is

If a shootout were actually a crapshoot, you'd expect the save percentages of goalies in the shootout to cluster around the same point, and that is not the case.

Those articles you linked to are showing a failure of deductive logic. They take as fact "there is not enough evidence to show that skill affects shootout results" and then somehow conclude from that that "the shootout is a crapshoot." The correct conclusion would be "we don't know if the shootout is a crapshoot." That's assuming that the original premise "there is not enough evidence to show that skill affects shootout results" is correct, and I would say that variations in save percentage in the shootout show that is likely incorrect.

 

[Evincar]

Assists are the most overly weighted, inaccurate abstraction of a statistic. The mere existence of the assist defies basic logic. What is being counted: shots or passes? Why is the assist worth an individual point equal to the value of a goal? Why are there two assists awarded? Why is setting up a goal worth twice the value of the goal itself?

When one reads a player's assist totals, what does it actually say about his game? Breaking down between first & second assists still tells us nothing about whether someone is shoot first or pass first, & it tells us virtually nothing about a person's shooting or passing ability other than the fact that they touched the puck shortly before their teammate scored. Period.

Get rid of assists & points & replace them with more accurate stats, like maybe

* completed passes
* shot assists, or shots leading to goals [ie deflected or tipped in]

Worst idea ever. Assists can be just as valuable as goals.

 

[matnor]

If a shootout were actually a crapshoot, you'd expect the save percentages of goalies in the shootout to cluster around the same point, and that is not the case.

On the contrary, that is exactly what the the first figure in each of the two cited articles show. If shootouts are truly random then we would expect 99% of the goalies to lie within the confidence interval. The data does not seem to contradict this.

Those articles you linked to are showing a failure of deductive logic. They take as fact "there is not enough evidence to show that skill affects shootout results" and then somehow conclude from that that "the shootout is a crapshoot." The correct conclusion would be "we don't know if the shootout is a crapshoot." That's assuming that the original premise "there is not enough evidence to show that skill affects shootout results" is correct, and I would say that variations in save percentage in the shootout show that is likely incorrect.

I agree that it would be wrong to say that shootouts are crapshoots based on the data in the articles. Rather what can be said is that based on the data we cannot reject the null hypothesis that shootouts are crapshoots. But the authors seem to be aware of this. The first article says

Again this does not suggest that there is not skill differences between players (in this case goalies) but it does mean that the randomness (or noise) of the shootout is completely dominating the skill of the players.

and the second says

we still can't feel very confident that this is anything more than random chance, but the fact that more goalies are moving towards the lines as the dataset grows larger does suggest that there may be a talent here that we just haven't been able to measure yet

Fwiw, I believe there is reason to believe there is skill differences in shootouts but the articles do a fairly good job of showing that the difference is not big enough to be detectable (yet).

 

[TheDevilMadeMe]

On the contrary, that is exactly what the the first figure in each of the two cited articles show. If shootouts are truly random then we would expect 99% of the goalies to lie within the confidence interval. The data does not seem to contradict this.



I agree that it would be wrong to say that shootouts are crapshoots based on the data in the articles. Rather what can be said is that based on the data we cannot reject the null hypothesis that shootouts are crapshoots. But the authors seem to be aware of this. The first article says



and the second says



Fwiw, I believe there is reason to believe there is skill differences in shootouts but the articles do a fairly good job of showing that the difference is not big enough to be detectable (yet).

See these posts from this thread

Something else to consider when judging current goalies is the shootout. It has been part of the NHL since Lundqvist joined the NHL 8 years ago and can play a key role in a teams record.

During that time the Rangers goaltenders have the best Sv% at .738%. Lundqvist has a record of 45 W and 30 L, he has faced 287 shots and has a .763 Sv%. His backups have a 9W and 10L record and .631 Sv% on 65 shots.

Rangers shooters have only had a 31.2 shooting % in the shootout which is 22nd best.

In only 2 of the 8 seasons have the Rangers been more than 5 points from missing the playoffs so obviously the 45 extra points Lundqvist has helped get the Rangers is pretty significant.

In the last 8 years the Rangers late in games and in OT are definitely guilty of sitting back and playing for the shootout believing that even with NY's anemic offense their best chance to win is sending Lundvist 1 on 1 with the other teams best shooters.

Shootouts are an interesting aspect. They most definitely contribute to a team's overall success, but for both skaters and goalies they have been almost completely ignored when it comes to awards voting/discussion.

I think Lundqvist absolutely merits additional consideration based on his shootout performance. As much as people might not like it, it's a part of the game now.

Puck prospectus calculates the influence of shootout performance into GVT, so one can get an idea of what kind of impact a player can have. In 2011/12 for example, Kovalchuk contributed 5.2 goals above replacement, and the highest goalie, Varlamov, contributed 5.1. Definitely not negligible, even if year-to-year variance will naturally result in peaks and troughs. They estimate over the past two seasons (pro-rated for 82 games in 2012/13's case) Lundqvist contributed four goals above replacement. Not a huge amount, but a sizeable chunk (~7-8%) of his overall performance.

In larger samples the effects can be largely obscured, but in individual seasons players can really augment (or badly harm) their team's record in the shootout.

The article that calls shootouts a crapshoot is from 2010. Since then we've had 3 more seasons of sample size.

I think luck is a big component of the shootout, but I just find it incredibly hard to believe that it's the only component as is claimed by a "crapshoot." In general, I'm highly skeptical when someone makes the claim that a result is based off nothing but "luck."

 

[PSGJ]

It's easily goalie wins that is the most useless commonly used stat.

 

[Psycho Papa Joe]

Due to the wide variation in interpretation, hits, giveaways and takeaways. Only somewhat usefull when assessing players against their own teammates.