Most overrated and misapplied stats

[ScottM]

It always gets me that there are certain stats that people use in rating or comparing players that are given much more weight than they should be if they should be given any weight at all. This thread is going to be my soapbox to complain about some of those, and I invite you to do the same. I've got three stats I'm going to harp on here, and everyone is welcome to throw their own thoughts about these or any others into the ring.

 

1) Plus/Minus

I think plus/minus can have some value when taken in the proper context, but it's almost never taken in the proper context and almost always given too much weight. Too often, it seems that it's given as the stat that shows how good a player was defensively. The first problem with that is that's not its intended purpose. It's supposed to demonstrate the players' overall effectiveness on the ice by showing his goal differential. It can be telling on the surface. When Bobby Orr posted a +124 season, that was very telling, as was Bill Mikkelson's -82 season. But most of the time, to have any value, it has to be taken in context of the team. For example, one guy finishes the season -5 and another gut finishes +10. Which one is better? Well, if the first guy's team got outscored by 80 goals and the other guy's team outscored their opponents by 150 goals, it seems fair to say that the first guy did better despite the "worse" plus/minus rating.

 

2) Game-Winning Goals

I think this one is absolutely worthless. Now, if we're talking about an overtime winner, and we see a guy that seems to collect a lot of those, we might be on to something. In that case, it could well tell us how a guy responds to pressure. As a rule though, I think this stat tells us a grand total of nothing. Basically, they guy that scores the "game-winner" is the guy that just happens to score his goal at the right spot in the game. Let me give a couple of examples of what I mean. First, a team wins a game 8-1, and the winning team's second goal is scored five minutes into the first period. What does that tell us? Well, nothing, other than the fact that that guy scored a goal early in a game before it got completely out of hand for the other team. Second, the team wins 5-4. It's a closer game, so that means the game-winner means more, right? Not necessarily. What if they took a 5-0 lead, only to collapse in the third period to nearly blow the lead. In this case, the goal was definitely needed for the win, but what makes it in particular more valuable than any of the other four? Nothing, I'd argue.

 

3) Wins (Goalie Stat)

Ultimately, it's all about wins as a team, but I think that the stat of wins for goalies is grossly overvalued. If there's any stat that is a complete team effort, this is it. Now, to be sure, there are some games that a goalie stands on his head and seems to win nearly singlehandedly, but that's not always (not usually) the case. If a team wins 6-5, it seems pretty obvious that generally indicates that the offense had a lot more to do with the win than the goaltending, and yet that goalie that had a pretty miserable game, giving up five goals, is credited with the win. What about a 2-0 game in which the goalie only faces 10 shots? Give him credit for stopping everything he faced, but don't you think the defense in front of him deserves as much credit (if not more) for that win. Had he faced 40 shots, it's very likely that the outcome would have been different. In short, all wins for goalies aren't created equal, and to me, this stat creates the illusion that they are.

[Podein25]

@ScottM

 

Good topic.

 

I don't really disagree with anything you wrote. I have strong opinions on the plus/minus thing because, as you say, people often misunderstand it's purpose. For me, it's a decent and reliable barometer of a player's even strength play.

 

Detractors often point to the cases where bad luck or good luck amounts to a minus or a plus, thereby rendering the stat "useless," but in truth those are exceptions. I'm biased, because as a player I always kept track of it and was proud of being a plus player. Because it meant you won more battles than you lost, you won shifts, you put the puck in the net and kept it out consistently. The coach puts those plus players on the ice when he needs a goal and when he needs to defend a lead.

[AJgoal]

GAA for a goalie is almost as useless as wins. It's also a measure of the team's defense. Poor defense usually equals a poor GAA.

 

While save percentage tells a bit more of the story, it can also be misleading. A goalie that faces a low number of shots on high quality chances is likely to have a lower save percentage than someone who faces a ton of perimeter shots. Still it grades out better on the goalie's performance.

[yave1964]

Plus/minus of course is a biggie, like you both point out, it serves a purpose but it is generally misused.

 

GWG, another really good one. No argument.

 

 steals, not a super common one but still a league used stat is incredibly misleading, I remember a game last year against the Leafs where Datsyuk was all over the ice taking pucks away almost at will and the box score that night he was credited with one steal, he seemed to have one per shift. What constitutes a steal?

 

 Shots on goal, IMHO should be redefined, lots of time a goalie faces 25 shots in a night but the whole story is the five posts that didn't count, the fifteen blocks by the defenders. I think that shots on goal and shots on net are two separate things and could be easily tracked.

 

  Assists are one of the most abused stats around, more often than not two assists will almost automatically be handed out to the home team on a goal unless their is incredible evidence to the contrary whereas the road club not so much. It has admittedly gotten somewhat better over the years but not evened up by a damn shot.

 

 

[TropicalFruitGirl26]

@ScottM

 

Can't say that I can argue with any of your points on the stats mentioned.

Generally speaking, it's basically how I feel about them.

 

Some extra commentary though on some things:

1---Plus / Minus

While I agree it isn't all telling and many times too much weight IS indeed put on it, it certainly works well in conjunction with other stats, such as TOI and team record.

Why?

 

Well, because if a guy spends a lot of time on the ice, and sports a particularly low or high +/-, I think its a fair indicator of how his presence on the ice is affecting goals being scored for or against because he is on there the majority of the game.

He either is a big part of the goals for, or one of the big reasons teams are getting (and burying) good chances.

 

Of course, that too is NOT all encompassing. If a team has a leaky goaltender, it is hard to fault a skater if the goalie simply isn't making the stops.

Which leads to overall team record.

On a bad team, expect even GOOD players to have putrid +/- ratings, on a good team, a pedestrian player can have an All Star-like +/- number.

 

Heck, a guy who is -2 on a horrible defensive team that is 15-32-5 can be said to be doing quite well as opposed to the guy who is +10 on a team that scores a ton (many of which may have nothing to do with him) and is likely something like 25-12-5, just for an example.

 

Also, one last point about +/-, I think POSITION should factor into how much weight it carries.

What I mean is, I would personally put more weight on a +/- stat for a defenseman over a forward.

Not saying it should be the way to evaluate a d-man, but let's face it...a d-man's job is to, well, DEFEND...and part of that is to help in goal prevention.

 

Repeating myself, it is NOT all telling, all encompassing, even for D-men, but since a forward's job is to put pressure and generate offense moreso than a defenseman, to me, +/- has much less value for a forward than a D-man.

 

So plus/minus, taken with a grain of salt can be useful, but it is HARDLY the tell all some people make it out to be.

 

2---Goalie Wins.

Again, by itself, doesn't tell you much about the actual goaltender.

This stat, IMO, HAS to paired with sv%, GAA, and even high/low danger shots faced to see whether the goalie gets wins on "easy chances" or is put through the wringer to get his W's.

 

Last season, Cory Schneider and Ben Bishop ranked pretty high in the Sv% and GAA rankings (Bishop led the league in GAA and was second in %, while Schneider was 4th in both GAA and %, on a bad team no less), but BOTH goalies were considered "inferior" to the kind of season that a guy like Braden Holtby had (even though the GAA and % numbers of both Schneider and Bishop were superior), simply due to the fact that Holtby had a whopping 48 wins, compared to Bishop's "ho hum" 35, and Schneider's "putrid" 27.

 

Was Holtby a good goalie last season? Absolutely. Was he THAT much better than either Schneider or Bishop? Not necessarily...and a case can be argued (not here though!) that either Bishop or Schneider were a tad BETTER last season.

 

Someone's gotta get the wins and losses recorded to them right? Well, the goalies are the guys. Just like pitchers in baseball. But just like MLB hurlers, the WINS are not the most telling stat if taken just by itself.

As Scott said, context is oh so important.

 

3--Game winning goals.

Yes, absolutely useless. No ifs ands or buts about it.

The just HAPPENS to have scored the last goal when the opposition STOPPED scoring theirs. How very fortunate for him.

 

Guy gets 10 GWG's in 82 games.

Awesome right? Well, not really. Not if that's ALL he managed the entire year and was just lucky enough to have scored his once the other team stopped.

Extreme example, yes, I know, but I think the point is made.

 

Other than for fantasy value, GWG is not something I EVER look at to see if a player has "game"...especially when there are much better measuring sticks available.

[yave1964]

Here is one I would like to see:

 

Shootout GWG counting as a goal. Not every shooter, just the one who scores the winner I would like to see it count as a goal and a GWG.

[AJgoal]

Go ahead goals are a much better stat for performance than GWG. I find GTG to be a good one, too. 

 

But yeah, Guy could score the 5th goal of a game that starts as a blowout, then his team relaxes in the third and the opponents score 4. That 5th goal was nothing when it was scored, but in the end stands up as the game-winner. Your team just happened to not stink quite enough to let them all the way back into it.

 

Totally unrelated, but my team is utterly incapable of letting me ever have a goal stand up as the game-winner.

[JR Ewing]

1 hour ago, TropicalFruitGirl26 said:

Also, one last point about +/-, I think POSITION should factor into how much weight it carries.

What I mean is, I would personally put more weight on a +/- stat for a defenseman over a forward.

Not saying it should be the way to evaluate a d-man, but let's face it...a d-man's job is to, well, DEFEND...and part of that is to help in goal prevention.

 

Repeating myself, it is NOT all telling, all encompassing, even for D-men, but since a forward's job is to put pressure and generate offense moreso than a defenseman, to me, +/- has much less value for a forward than a D-man.

 

 

I hold a mostly opposing viewpoint: +/- is not necessarily a great indicator of a defenseman's ability to play in his own end, because it doesn't credit work in that end with a plus. That comes from goal production. For me, the relative lack of minuses, or chances surrendered, is the key to blueliners.

 

But all of that is fine, because what IS defense?

 

Is it hitting?

Is it boxing out in the home plate area?

Is it breaking up cross-ice passes?

Is it punishing players in front of the net?

Is it breaking up the cycle?

Etc, etc, etc.

 

My favorite viewpoint on this comes from Daryl Sutter:

 

“The big thing in today’s game is you have to be able forecheck and backcheck, and you have to have the puck. You can’t give the puck up. We don’t play in our zone, so there’s not much defending. The game’s changed. They think there’s defending in today’s game. Nah, it’s how much you have the puck. Teams that play around in their own zone think they’re defending but they’re generally getting scored on or taking face-offs and they need a goalie to stand on his head if that’s the way they play.”

 

It's good to know how to do those things above once the puck is in your end, but the best defense happens at the other end of the rink, when you're not giving up chances. With regard to defensemen, plus/minus is a blunt instrument.

 

 

[WordsOfWisdom]

2 hours ago, ScottM said:

1) Plus/Minus

I think plus/minus can have some value when taken in the proper context, but it's almost never taken in the proper context and almost always given too much weight. Too often, it seems that it's given as the stat that shows how good a player was defensively. The first problem with that is that's not its intended purpose. It's supposed to demonstrate the players' overall effectiveness on the ice by showing his goal differential. It can be telling on the surface. When Bobby Orr posted a +124 season, that was very telling, as was Bill Mikkelson's -82 season. But most of the time, to have any value, it has to be taken in context of the team. For example, one guy finishes the season -5 and another gut finishes +10. Which one is better? Well, if the first guy's team got outscored by 80 goals and the other guy's team outscored their opponents by 150 goals, it seems fair to say that the first guy did better despite the "worse" plus/minus rating.

 

Hit the nail on the head. It's a measure of goal differential, not defence. +/- was the first attempt to create a single statistic that measured the overall value of a hockey player. The highest "+" player was intended to be the best overall player, period. 

 

Conclusions one can draw from +/- include:

• Higher is better.
• The stat is team-based but also line-based. It's best to compare players in +/- within their own team.
• If a player scores 82 points (1 PPG) and still has a negative +/-, that player sucks defensively. 
• If you want to extract the DEFENCE (the "-") from +/-, use the following formula: (+/-) - (PTS - PPP)

 

 

 

[ScottM]

@AJgoal For the reasons you mentioned, I always like to consider goals-against average, save percentage together, and shots faced together. The more information you get, the better your chance to get the full picture. Sometimes, taking it all together helps to cut through some of the factors in your post that one alone can't do.

 

@yave1964 I really like your idea of separate stats for shots and shots on goal. It wouldn't at all be hard to do. Next Saturday, high school soccer season will start in Mississippi, and two of the stats I'll be collecting at the games I cover are shots and shots on target. Now, maybe it's not at as fast of a speed as hockey, but still, it's an easy thing to do.

 

Assists... At this point, I don't think things are too bad. As you said, they're getting better about that, but I'd argue that the current rules governing them are somewhat faulty. A guy just barely taps the puck in his defensive zone, one of his teammates takes possession, makes a complete individual play, and the first guy gets an assist. Why? What did he do to set up the goal? On the flip side of that, I saw a video of a goal recently in which a defenseman made an outstanding breakout pass, the guy he passed it to advanced it through the neutral zone, that guy took it in, passed it forward to a guy near the net, and he pushed it across in front of the goalie to set up a goal. It actually makes more sense to give four assists on that goal than it does to give even one on the first one. I'm not suggesting that we start giving four assists per goal, but I think there should be more common sense rules for it. You might have to let Toronto govern it to make sure it's done consistently, but I think it would be a good step.

 

I have mixed thoughts on the shootout goal. I keep going back and forth on that. Either way, I still hate that stupid shootout, for whatever that's worth.

 

@TropicalFruitGirl26 I think everyone here knows that I think the voters completely blew the Norris last year, many probably know that I think they blew the Conn Smythe, but what might not be as well known is that I think they blew the Vezina. I fully believe that the only reason Holtby won was an overestimation of the wins stat. The fact that he broke the season wins record tells you two things. 1) He played a lot of games (which is something deserving of credit) and 2) he played on a really good team (something he's only a part of). I felt strongly that both Bishop and Schneider were more deserving of the trophy (Bishop was my pick). It goes back to taking all of the information at our disposal and properly applying it.

 

@JR Ewing It's funny how that quote keeps coming up, isn't it? Is it a good thing for a defenseman to do the things you listed? Absolutely. It's good for any player to be able to do them. But, the question of "what is defense?" is a valid one here. Defense, in my mind isn't something that can accurately be defined statistically, so I think it's best to stop trying. Any player should be measured by their all around contributions to their team. All of the stats mentioned help paint that picture, but none of them come close to doing it alone.

[WordsOfWisdom]

16 hours ago, yave1964 said:

Shots on goal, IMHO should be redefined, lots of time a goalie faces 25 shots in a night but the whole story is the five posts that didn't count, the fifteen blocks by the defenders. I think that shots on goal and shots on net are two separate things and could be easily tracked.

 

The shots statistic that is listed with player stats is "shots on goal".  It is not a shot on goal if it gets blocked or misses the net.

 

[yave1964]

2 hours ago, WordsOfWisdom said:

 

The shots statistic that is listed with player stats is "shots on goal".  It is not a shot on goal if it gets blocked or misses the net.

 

Correct, but my point is that I think that it leaves a lot to be desired. I believe that is an accurate way to judge a goalie but not a shooter.

  As an example, using Holland and Komarov from last season, Komarov in 67 games got 138 shots on net and Holland in 65 games managed 130. Sounds like they were pretty even right?

  But what is missing from the equation is how many times did each of them fire and miss wide or have the shot blocked? I have to believe this would be fairly easy to track, if you can track how many shots a defenseman blocks you should be able to judge how many shots a forward has blocked or misses wide.

  Why is that important? IMHO a guy like Komarov with a 14 percent shooting percentage, not too shabby really, may have had fifty blocks/net misses throughout the year, lowering it significantly whereas a Holland may have had ten or so meaning that his overall pct. is in fact much closer to what Komarov's was. IMHO it would give a true rating for a shooter as to his overall effectiveness. Shots that made it through by a shooter do indeed tell part of the story but not the whole thing.

  An example would be Jordin Tootoo, in 2013-14 with the Wings he managed 3 goals on 45 shots, a poor but somewhat acceptable level for a 4th liner, but the eyeball test, he consistently missed wide and to the right by five freaking feet more than anyone I have ever seen, it seemed once a game he would wind up and my jaw would clench knowing that if the other team would simply allow the Wing sto move the net five feet over that Tootoo would be a 30 goal a year guy and invariably uncontested he would hit the glass with a nifty thirty footer. If you add in the 30 or more times that he did this it makes playing him unacceptable. Same thing with Teemu Pulkinnen last season, he has a HUGE slapper, which works in the AHL but in the NHL where d-men are a bit quicker he gets them blocked or he would rush and miss wide with alarming regularity.

  So shots, not just shots on goal IMHO would be the true benchmark of how to look at a shooter, the system should be subtle enough to differentiate.

  And to finish, lets take a look at Kadri who led your team in shots last year by a wide margin, my guess would be that when he shoots he almost never misses the net, they may not always be the best scoring chance but they almost always get through and if they do not go in he can at least hope for a juicy rebound. Contrary JVR who missed half the season averaged the same amount of shots per game but judging stylistically between them using the eyeball test, I would bet that in half the games JVR probably had as many misses or blocks as Kadri did in a full season. That is the problem with what we have now, it only defines how many get through and does not give you a real read on the shooters actual ability.

[TropicalFruitGirl26]

@yave1964

 

Good post.

SOG by themselves don,t tell the whole story.

That said, I believe shots attempted, what you seemed to be alluding to IS a tracked stat, albeit not as widely used or acknowledged as much as pure SOG.

 

Now within those 'shots attempted' is there a differentiation made between those blocked or missed? Not sure.  I would have to dig further into that.

 

I have seen shots missed being tracked on some sites....but not quite sure what defines a missed shot. Are blocked shots missed attempts or do they mean just the ones actually missing the goalie?

 

Not to mention hit posts and crossbars, while technically hitting the net is not a SOG because the goalie never touched it......though oftentimes that post or crossbar is a shot that flat out 'beat' said goaltender.

 

So posts / crossbars count the same as that Tootoo shot that hit the glass 10 ft from the goalie.....though EVERYONE knows the post hitter is much more accurate and likely a better producer on the season.

 

But yea, SOG by itself is often grossly misused.

[SpikeDDS]

@yave1964

@TropicalFruitGirl26

 

I was always under the impression that a post or crossbar shot was NOT a SOG. I know that is the way it is in soccer. Is it different for hockey? I didn't think it was. For it to be a shot on goal, it has to actually be going in the goal if the goalie didn't stop it. Right?

[yave1964]

1 minute ago, SpikeDDS said:

@yave1964

@TropicalFruitGirl26

 

I was always under the impression that a post or crossbar shot was NOT a SOG. I know that is the way it is in soccer. Is it different for hockey? I didn't think it was.

Correct, a post or a crossbar is not a SOG. But it should be tracked for the shooter to get a better read on his real overall accuracy.

[TropicalFruitGirl26]

1 minute ago, SpikeDDS said:

@yave1964

@TropicalFruitGirl26

 

I was always under the impression that a post or crossbar shot was NOT a SOG. I know that is the way it is in soccer. Is it different for hockey? I didn't think it was.

 

it isn't a SOG.

That was one of the points I was making. As far as SOG, it never happened.

 

Thus like what Yave and I were saying the SOG wouldn't tell you that part of the story

[WordsOfWisdom]

32 minutes ago, yave1964 said:

But what is missing from the equation is how many times did each of them fire and miss wide or have the shot blocked?

 

Ah good point. The NHL should implement a total shots stat and an "overall shooting %" to measure goals on total shots.

 

[SpikeDDS]

Anyone like Corsi? I always thought it was a relatively revealing stat, even if not a generally published one.

[WordsOfWisdom]

4 minutes ago, WordsOfWisdom said:

 

Ah good point. The NHL should implement a total shots stat and an "overall shooting %" to measure goals on total shots.

 

 

Also, I know this information is tracked as part of the NHL's extended stats. 

 

For each player they have total shots, missed shots (a post also being a miss), blocked shots, and SOG. 

 

SOG + missed + blocked = total shots

 

Either an overall shooting percentage (which would probably be quite low, somewhere around 5% for a typical player) or a percentage of total shots that become SOG would be useful.

 

Overall shooting % = goals / total shots

 

Shot efficiency % = SOG / total shots

 

[SpikeDDS]

The thing about most stats is they are almost completely if not completely devoid of context.  That's why I kinda like the more contrived stats like Corsi and others, because they at least TRY to give SOME context to a number.

[WordsOfWisdom]

3 minutes ago, SpikeDDS said:

Anyone like Corsi? I always thought it was a relatively revealing stat, even if not a generally published one.

 

It's okay. I think of it as the +/- for SOG. 

[yave1964]

29 minutes ago, SpikeDDS said:

The thing about most stats is they are almost completely if not completely devoid of context.  That's why I kinda like the more contrived stats like Corsi and others, because they at least TRY to give SOME context to a number.

Corsi is gaining in popularity. I like anything that gives more of a true read as to how a player is doing.

[ScottM]

I am a fan of Corsi. I think it, like plus/minus, has to be taken in context, but I find it more useful overall. I mention the context because a guy with a CF% of 51 on a team with a CF% of 47 is doing a lot better than a guy with a 51 on a team with a rating of 53.

 

As for the shots on goal vs shots thing goes, I'll tell you how I look at the concept while statting soccer games. Imagine there's no goalie. If the shot would have gone in without the goalie, it's a shot on target. If it fails to go in goalie or not, it's just a shot. Like I said before, I see no reason the separation couldn't be made. It's really not a hard thing to keep up with.

[WordsOfWisdom]

55 minutes ago, ScottM said:

As for the shots on goal vs shots thing goes, I'll tell you how I look at the concept while statting soccer games. Imagine there's no goalie. If the shot would have gone in without the goalie, it's a shot on target. If it fails to go in goalie or not, it's just a shot. Like I said before, I see no reason the separation couldn't be made. It's really not a hard thing to keep up with.

 

Exactly. The NHL tracks all of that, but they don't release it to the public as far as I know.  

 

If you took a player like Ovechkin for example, his stat line might be something like this:

 

50 G, 398 SOG, 124 missed shots, 111 blocked shots, 633 total shots

 

His shooting % (SOG only) would be 50 / 398 = 12.6%

His overall shooting % (on all shots) would be 50 / 633 = 7.9%

His shot efficiency % (shots -> SOG) would be 398 / 633 = 62.9%

 

[ScottM]

2 minutes ago, WordsOfWisdom said:

 

Exactly. The NHL tracks all of that, but they don't release it to the public as far as I know.  

 

If you took a player like Ovechkin for example, his stat line might be something like this:

 

50 G, 398 SOG, 124 missed shots, 111 blocked shots, 633 total shots

 

His shooting % (SOG only) would be 50 / 398 = 12.6%

His overall shooting % (on all shots) would be 50 / 633 = 7.9%

His shot efficiency % (shots -> SOG) would be 398 / 633 = 62.9%

 

Right. And something like that is far more telling to me than a simple shots statistic. Just think if we had that throughout league history. It would certainly help in comparing goalscorers.