TSN: Jake Gardiner or: How I Learned to Stop Worrying and Love the Advanced Stats

[Swayze*]

I think what it means (correct me if I'm wrong) is that all the goals against that were scored against him tonight, regardless of whether he was out of position, afraid to lay a hit, WHATEVER,....

It was basically everyone else's fault, not his fault, just a few simple Corsi events and not statistically meaningful, and he had a GREAT GAME overall and (insert these numbers from this website) here's why!!!

GO JAKE!!!!

Well he was -2. Made some srs blunders

I thought he had a poorish game but I will honestly defer to the advanced stats group

Im open to it

 

[Bomber0104]

Sure, they're treating every shot from that particular area as equal, but they're based on the aggregate shooting-percentage from that area so 'screened shots' are inherently accounted for, they're just not differentiated.

Why would you treat a spam shot as one of quality?

Why would you treat a shot that a goalie can see the same as one that he can't?

Why would you treat a shot that a goalie can set up for the same way as one where players are purposefully trying to make him move, so as to not?

Also, if we're assuming that screened shots are causing more goals (on either side), then a player allowing a higher frequency of screened shots would have a higher GA60.

If that isn't the case, then it doesn't matter, because screened/'unscreened' shots have little correlation with goals allowed (and, in turn, little correlation with Win%).

This isn't encapsulated in your statistic, though.

You would never know whether the team is pinned in their own end on their heels, and on the receiving end of a heavy point shot with traffic (a very negative event, with a high percentage of scoring) or whether a player is coming in 1-on-2 and spamming a shot at a goalie while his linemates are changing (a common, tactile event with next to zero probability of scoring).

If you want to value/devalue the players in both of these scenarios in symmetry, be my guest, but they are in no way symmetrical and it's completely lazy for whoever made that website (that you are pawning these "stats" from) to suggest

So, assuming screens/angle/traffic etc. makes a difference,

There is no assumption to make. A point shot that a goalie can see will be saved with ease vs. a shot with traffic (which reduces a goalies visibility and increases the chance of a deflection).

doesn't your claim that Gardiner is always out of position- and never doing anything well- mean that Gardiner would allow more goals against?

I've not suggested once that Gardiner is guilty of this....

Rather I'm highlighting the absolute gaping holes in the logic behind these little spatial datasets that you are incorrectly interpreting.

If so, either you're doing very poor job of analyzing what is really happening (very, very likely), or these variables that you assume make a difference don't really make a difference at all, because Gardiner is not allowing a bunch of goals against.

So which is it?

Again, I was using one simple example highlighting the folly of this internet stat you're using.

Go back to my first paragraph at the top of my post for more to ponder over.

 

[Bomber0104]

Gardiner's bad and the Corsi ghost is helping him. It's the only way people like him will accept the validity of advanced stats when they've already made up their mind about a player. An argument that will change his mind about Gardiner doesn't exist.

The problem with Corsi, inherently, is that it neither appropriately apportions credit / blame to an individual player, nor does it recognize shot quality / passing quality/ defensive quality.

Imagine a workplace where large successes or failures were recognized in the same way as minor ones? And rather than you getting the credit or blame, everyone does equally!

Is that a fair system?

 

[Swayze*]

The problem with Corsi, inherently, is that it neither appropriately apportions credit / blame to an individual player, nor does it recognize shot quality / passing quality/ defensive quality.

Imagine a workplace where large successes or failures were recognized in the same way as minor ones? And rather than you getting the credit or blame, everyone does equally!

Is that a fair system?

Same as +-

 

[The Thin White Duke]

The problem with Corsi, inherently, is that it neither appropriately apportions credit / blame to an individual player, nor does it recognize shot quality / passing quality/ defensive quality.

Imagine a workplace where large successes or failures were recognized in the same way as minor ones? And rather than you getting the credit or blame, everyone does equally!

Is that a fair system?

Ok except this workplace does thousands upon thousands of projects per year, and it's noticed that when one of the employees is present the overwhelming trend is that projects get done faster/more successfully. None of his peers demonstrate a similar effect even when we account for the difficulty of the project. It doesn't matter whether this employee is truly contributing at a higher rate, or he's making constant mistakes and the Project Ghost following him is somehow speeding up the groups cumulative performance, when he's there things just get done. If your concern is net progress, this employee is objectively a good one to have, no?

 

[Bomber0104]

Same as +-

Exactly...

They are BOTH differentials.

If you are going to crap on +/- , well then I hope you hold Corsi to the same standard because that, too, is a simple differential.

Corsi fanatics neglect this simple fact but because they have more events, as they say, so they think they are held to a different standard.

The object of the game ultimately is to out-score the opponent however, and therefore you are punishing players who can pull off the high-end offensive and defensive maneuvers by holding shots to an undeserved standard than scoring and preventing goals.

 

[Bomber0104]

Ok except this workplace does thousands upon thousands of projects per year, and it's noticed that when one of the employees is present the overwhelming trend is that projects get done faster/more successfully. None of his peers demonstrate a similar effect even when we account for the difficulty of the project. It doesn't matter whether this employee is truly contributing at a higher rate, or he's making constant mistakes and the Project Ghost following him is somehow speeding up the groups cumulative performance, when he's there things just get done. If your concern is net progress, this employee is objectively a good one to have, no?

In response to this, I'll reiterate my position and say attempted projects that cannot be accredited to any individual(s) neither in quality or quantity cannot be held as conclusive evidence of performance.

After all, shot-differential numbers are in-and-of-themselves proxy measurements which don't perfectly correlate to possession.

I wouldn't feel confident in making grandiose claims about a player's ability based on already sketchy data.

Maybe this is why NHL teams are actually beginning to use spatial and velocity models to track puck and player movements?

A good friend of mine (a UWaterloo grad. engineer) is working on some stuff for a few NBA teams, on behalf of his company. Incredible really. All done with cameras, logarithms, and precisely tuned models with input from the teams (who understand the sport) that are buying their product.

In time you guys will realize how novice this crap is and finally move on.

 

[The Thin White Duke]

In response to this, I'll reiterate my position and say attempted projects that cannot be accredited to any individual(s) neither in quality or quantity cannot be held as conclusive evidence of performance.

After all, shot-differential numbers are in-and-of-themselves proxy measurements which don't perfectly correlate to possession.

I wouldn't feel confident in making grandiose claims about a player's ability based on already sketchy data.

Maybe this is why NHL teams are actually beginning to use spatial and velocity models to track puck and player movements?

A good friend of mine (a UWaterloo grad. engineer) is working on some stuff for a few NBA teams, on behalf of his company. Incredible really. All done with cameras, logarithms, and precisely tuned models with input from the teams (who understand the sport) that are buying their product.

In time you guys will realize how novice this crap is and finally move on.

Ok lets just look at the numbers on their own, without any ties to on ice possession. We have two options: Gardiner is doing something to deviate these numbers from the norm, OR, Gardiner is consistently "lucky" in a specific direction for lack of a better word. Which of these is more likely to you on a several-thousand repetition basis? I'm going to assume you'd agree that he's doing something to impact these numbers.

This leaves us disagreeing on what a change in those numbers means yes? So our roadblock here is that you don't agree that in the aggregate a better ratio of shots for vs against correlates to a better ratio of goals for vs against. Or is it that you can't pull a single player's impact out of 5-man unit statistics?

If it's the first roadblock, then there's no way we can find a middle ground. If it's the second, you must believe in the corsi-ghost following Gardiner which accounts for the disparity between how that unit preforms with and without him.

Better tracking software/hardware is absolutely welcome, our current stats are only really useful in the aggregate. Better tracking will be great for in-game and short term use. You're going to be disappointed if you're expecting it to contradict existing stats in the long term though.

 

[91Kadri91*]

Why would you treat a spam shot as one of quality?

A spam shot? The more shots you attempt, the more often you will score.

That's not to say that all shots are equal, but to deny the correlation between volume and total result makes you either: (a) a moron or (b) disingenuous.

Why would you treat a shot that a goalie can see the same as one that he can't?

They don't have the means to differentiate it (since NHL PbP doesn't track 'screen shots').

They're not assuming anything, they're simply calculating statistics (statistics they're able to calculate) that have a significantly greater reliability (more repeatable) than GF%- as well as being a significantly better predictor of future GF% than current GF% is- does, and significantly better correlation with Pt% than basic statistics (hits, takeaways, giveaways, penalties etc) have.

Why would you treat a shot that a goalie can set up for the same way as one where players are purposefully trying to make him move, so as to not?

Again, you're not addressing my point: if these arbitrary, hypothetical situations that you've conjured don't result in goals, how important are they? They'll be accounted for in scoring chances, they just won't be directly given a greater weight than a presumably less dangerous situation. Of course, if they do result in more goals against, the volume of these occurrences will be indirectly reflected in a player's GA60.

This isn't encapsulated in your statistic, though.

It's not my statistic, and no singular statistic encapsulates every situation and every result, which is why you use a myriad of statistics (which I have; read the entire thread).

Also, you've made no mention of xGoals, which does account for rebounds and shot angle (which has been mathematically proven to have an impact on Sh%).

You would never know whether the team is pinned in their own end on their heels,

Possession time would tell you that (and, by extension, Corsi).

and on the receiving end of a heavy point shot with traffic (a very negative event, with a high percentage of scoring)

If these events results in a high-percentage of goals (high Sh%), then these results (the proportion) will be (indirectly, but inherently) accounted for in the aggregate (not single game) GA60.

or whether a player is coming in 1-on-2 and spamming a shot at a goalie while his linemates are changing (a common, tactile event with next to zero probability of scoring).

If they're not going to result in goals (again, you're offering no proof for these assumptions, but I'll play along), then does it matter if you surrender profuse amounts of these events? These events have to result in a negative outcome (percentage based over a large sample) for them to be negative.

If you want to value/devalue the players in both of these scenarios in symmetry, be my guest, but they are in no way symmetrical and it's completely lazy for whoever made that website (that you are pawning these "stats" from) to suggest

They're not symmetrical, and no one is suggesting they are. What I'm stating is that a negative event (and the outcome of said event) will either be reflected in Corsi/Fenwick/Shot+Scoring Chance/High-Danger Scoring Chance or Corsi/Fenwick/Shot+Goal/Goal Against.

The only thing 'differentiating situations' could result in is the ability to better predict future GF% (and consequently future Win%), since there will never be a better indicator of Win% (that isn't Win%) than GF%.

They (situational differentiation statistics) won't do a better job of determining an individual player's current performance, they'll just help determine what kind of player said player is, and how they may (potentially) perform in the future (which seems to be of no interest to you).

There is no assumption to make. A point shot that a goalie can see will be saved with ease vs. a shot with traffic (which reduces a goalies visibility and increases the chance of a deflection).

Right, which will be inherently reflected in GA60 (over a large enough sample).

I've not suggested once that Gardiner is guilty of this....

Sure thing, kid.

I'll admit, all the goals against he's directly caused over the years have kind of bugged me

Gardiner completely blowing his coverage, letting his man walk in for a chance and goal to open the game.

And you're back to lying again.

Rather I'm highlighting the absolute gaping holes in the logic behind these little spatial datasets that you are incorrectly interpreting.

I'm not incorrectly interpreting them; I recognize them for what they are. They're the best predictive options currently available, but they're certainly not perfect, and I've never assumed them to be such.

Of course, even-strength goal differential has a near perfect correlation with Pt%, but I assumed your point was referring to Corsi/Fenwick/SC.

Again, I was using one simple example highlighting the folly of this internet stat you're using.

No, you created a hypothetical situation to argue a position I wasn't taking so that you could create an argument that you still can't win.

Go back to my first paragraph at the top of my post for more to ponder over.

Speaking of folly...

 

[91Kadri91*]

All done with cameras, logarithms, and precisely tuned models with input from the teams (who understand the sport) that are buying their product.

In time you guys will realize how novice this crap is and finally move on.

Yeah, like SportLogIQ:



But they're quite fond of Gardiner too, so they must be doing something wrong, amirite?

 

[Dustin]

Yeah, like SportLogIQ:



But they're quite fond of Gardiner too, so they must be doing something wrong, amirite?

Thanks to you and Nithoniniel for helping me understand how to interpret these stats.

Edit. Sentence structure.

 

[Bomber0104]

Ok lets just look at the numbers on their own, without any ties to on ice possession. We have two options: Gardiner is doing something to deviate these numbers from the norm, OR, Gardiner is consistently "lucky" in a specific direction for lack of a better word. Which of these is more likely to you on a several-thousand repetition basis? I'm going to assume you'd agree that he's doing something to impact these numbers.

He would be impacting them on a several thousand repetition basis. The significance and meaning of the underlying data (because that's all I am really willing to classify this stuff as) is questionable and is absolutely what I take issue with.

I'm interested in information, and information only.

Not data, no matter how much you toy with it.

This leaves us disagreeing on what a change in those numbers means yes?

Absolutely. I don't believe making inferences on an inference is a viable way to draw conclusions.

So our roadblock here is that you don't agree that in the aggregate a better ratio of shots for vs against correlates to a better ratio of goals for vs against.

This is generally how I set up a model of the game of hockey.

Possession < Shots < Scoring Chances < Goals For/Against

There are players that stimulate / inhibit certain aspects of either of these four generic categories.

The model you are setting up here is entirely omitting a key variable below:

Possession < Shots < Scoring Chances < Goals For/Against

It's already been shown that the relationship between shots and possession is somewhat suspect. And then you want to further dilute this shady relationship by smoothing out the numbers in two independent 5-man subsets?

Oh and it gets better, you ignore a key metric in shot quality which allows for the all important Scoring Chance variable...

Or is it that you can't pull a single player's impact out of 5-man unit statistics?

You may be able to infer a player's impact on shots, goal for/against, and to a much smaller degree, scoring chances.

What you can infer on the ability to score / prevent goals .... inconclusive.

If it's the first roadblock, then there's no way we can find a middle ground. If it's the second, you must believe in the corsi-ghost following Gardiner which accounts for the disparity between how that unit preforms with and without him.

Well it is actually both...and this is because you've created an incomplete model.

Nothing to do with me, you've just created a very scant system where insignificant data is held to a higher importance than it actually is worth...

Better tracking software/hardware is absolutely welcome, our current stats are only really useful in the aggregate. Better tracking will be great for in-game and short term use. You're going to be disappointed if you're expecting it to contradict existing stats in the long term though.

Actually shot based stats just scratches the surface.

A better understanding of the game would prevent you from making mistakes in statistical modelling.

From there technology only simplifies things.

 

[Ovate]

Yeah, like SportLogIQ:



But they're quite fond of Gardiner too, so they must be doing something wrong, amirite?

As an aside from talking about Gardiner, SportLogIQ are a bunch of hacks. The statistical side of their project is terrible, which they try to hide by keeping it largely secret, and then try to make up for it with flashy videos and charts that don't tell you anything.

I think Gardiner's great, but I wouldn't use SPortLogIQ as a source for anything.

 

[Bomber0104]

A spam shot? The more shots you attempt, the more often you will score.

You're not really not understanding what you've done wrong, eh?

You are attempting to convince people that two shots from the blue line:

1) a 1-on-2 shot where a player is just lobbing it on net to complete a line-change

and

2) a point shot with heavy traffic in front of the net

are one in the same, in terms of their likeliness to result in a goal (shooting percentage)

This is precisely what you have been suggesting this entire time and exactly what is inferred by that silly chart.

I'm sorry but this is just not how the game works in real-life.

That's not to say that all shots are equal, but to deny the correlation between volume and total result makes you either: (a) a moron or (b) disingenuous.

That's exactly what you are suggesting using this silly little chart of yours.

You're implying that since X and Y shots are both shot at Z location, X and Y are as likely to score as the other, based on the spatial proximity on the ice.

Maybe you don't understand what you are inferring by posting that chart, but that's exactly what you are doing.

You are treating all shots the same. And you are penalizing the players on the ice who scored the point shot and rewarding the player who spammed the shot in.

I won't even call you a moron for it, either. I'm completely tolerant and willing to help those who are uneducated in statistics and hockey.

They don't have the means to differentiate it (since NHL PbP doesn't track 'screen shots').

Well that's hardly my fault then is it, for disagreeing with you posting that chart here and worse, using data from it to make ridiculous claims?

They're not assuming anything, they're simply calculating statistics (statistics they're able to calculate) that have a significantly greater reliability (more repeatable) than GF%- as well as being a significantly better predictor of future GF% than current GF% is- does, and significantly better correlation with Pt% than basic statistics (hits, takeaways, giveaways, penalties etc) have.

That's the problem, dear. It's all done a bit too simply. There's no real method, or science involved, isn't there?

Sorry but I really don't have the time to address the rest of your post. It's all very repetitive and circular.

 

[champs*]

Go buy a ticket behind the leafs goal and just watch Gardner lol

 

[mikebel111*]

Go buy a ticket behind the leafs goal and just watch Gardner lol

I do and he is good

Hope he continues his fine season
Both he and Rielly looking great

 

[champs*]

I do and he is good

Hope he continues his fine season
Both he and Rielly looking great

He is not giving way in the corners when he hears footsteps lol? I haven't seen a game live this year

 

[91Kadri91*]

Every argument you made is dismantled by an aspect of my post that you chose not to reference.

You're not really not understanding what you've done wrong, eh?

You mean nothing? If I have done something wrong, I would appreciate it if someone would prove it. I'll be waiting.

You are attempting to convince people that two shots from the blue line:

1) a 1-on-2 shot where a player is just lobbing it on net to complete a line-change

and

2) a point shot with heavy traffic in front of the net

No, I'm not arguing any of that.

Here's what I'm arguing: more shots result in more goals. Here's what you're inferring: every shot is created equal.

are one in the same, in terms of their likeliness to result in a goal (shooting percentage)

I referenced scoring chances in the post you're responding to, and I explicitly discussed the differing shooting-percentages based on area; I am (clearly!) not arguing that every shot results in an identical Sh%. I'm well aware of the quantitative and qualitative difference between events (both the one's shown by scoring chances, and the one's not).

This is precisely what you have been suggesting this entire time and exactly what is inferred by that silly chart.

'That silly chart'? The scoring chances chart? First of all, the chart cannot infer anything- you do know that it's an inanimate object? It's a chart, Bomber... it's a CHART- and secondly, I haven't been suggesting anything of the sort. This is taken directly from the post you responded to:

They're not symmetrical, and no one is suggesting they are. What I'm stating is that a negative event (and the outcome of said event) will either be reflected in Corsi/Fenwick/Shot+Scoring Chance/High-Danger Scoring Chance or Corsi/Fenwick/Shot+Goal/Goal Against.

The only thing 'differentiating situations' could result in is the ability to better predict future GF% (and consequently future Win%), since there will never be a better indicator of Win% (that isn't Win%) than GF%.

They (situational differentiation statistics) won't do a better job of determining an individual player's current performance, they'll just help determine what kind of player said player is, and how they may (potentially) perform in the future (which seems to be of no interest to you).

I'm sorry but this is just not how the game works in real-life.

I'm well aware that there are nuances of the game that statistics don't capture, but they're nuances that no one can reasonable capture (unless you track every event in a game, which isn't publicly available data yet). They are also nuances that are reflected in aggregate goal for/against metrics.

You're inferring from my posts that I believe Corsi, Fenwick, Shots and Scoring Chances are the greatest statistical options that will ever be, even though I have not stated that in any of my posts. I have claimed that they're the best options we have (currently) for what they do (predict future GF; give us an idea of how a player will perform YoY), but they're far from perfect.

The problem here is that you're rebutting arguments that you've created.

That's exactly what you are suggesting using this silly little chart of yours.

You're implying that since X and Y shots are both shot at Z location, X and Y are as likely to score as the other, based on the spatial proximity on the ice.

No, I'm not implying that. The chart suggests that a shot from location X (based on proximity to the net) is more likely to enter the net than a shot from location Y; it makes no conclusion about different shots from a specific location. It doesn't even try to make a claim about different types of shots.

All it's saying is that a shot from location 'X' is more likely to go in than a shot from location 'Y', and they're right.

Maybe you don't understand what you are inferring by posting that chart, but that's exactly what you are doing.

I don't infer anything (I can't) by posting those charts. At best, I'm implying something. Of course, I didn't imply anything either.

You asked for the 'science' behind SC, and I provided it: the idea behind it is that shots from location X enter the net at a higher percentage than shots from location Y (which they do).

You are treating all shots the same. And you are penalizing the players on the ice who scored the point shot and rewarding the player who spammed the shot in.

Uh, no, the player who scored the point shot receives a GF60 (more indicative of current contributions to winning than any other individual statistic will ever be) and, at worst, a + Corsi, Fenwick and Shot. They could also receive a SCF if the shot was taken from 'centre-point'.

I won't even call you a moron for it, either. I'm completely tolerant and willing to help those who are uneducated in statistics and hockey.

That's good; you need a lot of help.

Well that's hardly my fault then is it, for disagreeing with you posting that chart here and worse, using data from it to make ridiculous claims?

What ridiculous claims did I make about it, exactly?

I said that Gardiner suppresses scoring chances (fact), that scoring chances (and corsi, and fenwick, and shots) is a better predictor of future GF than current GF is (also a fact), and that a puck from position 'X' on the ice will go in more often than a puck from position 'Y' in terms of percentage (also a fact).

These 'ridiculous' claims you're complaining about are claims conjured by you to create an argument nobody was, or wants to be, making.

That's the problem, dear. It's all done a bit too simply. There's no real method, or science involved, isn't there?

I'll just ignore the glaring double negative, and argue the point I know you were trying to make (pesky English language, amirite?).

There is a science behind it (I, and many others, have provided you with it), and it's really quite simple. Possession metrics like Fenwick and Corsi correlate (moderately) to future GF%, while current GF% does not. People use Scoring Chances, Corsi and Fenwick, because GF% does not help determine what future Win% may be; the sample of total goals is simply too small for there to be any YoY reliability and in-season predictive value. Essentially, possession metrics are a better indicator of what a player will be (and could be) than any statistic currently available (publicly).

The possession and shot statistics, meanwhile, are significantly more repeatable and predictive because the sample is significantly larger. They're the closest we come to possessing a perfectly valid (moderate correlation to Pt%) and reliable (moderate YoY) statistic.

Here's a link (which I already posted) showing the correlation between a myriad of statistics and Pt/Win%:

http://www.pensionplanpuppets.com/2013/7/10/4508094/what-statistics-are-meaningful-in-a-given-season-corsi-fenwick-PDO-hits-fights-blocked-shots

Sorry but I really don't have the time to address the rest of your post. It's all very repetitive and circular.

Sure thing, kid.

 

[CBinTokyo]

The problem with Corsi, inherently, is that it neither appropriately apportions credit / blame to an individual player, nor does it recognize shot quality / passing quality/ defensive quality.

Imagine a workplace where large successes or failures were recognized in the same way as minor ones? And rather than you getting the credit or blame, everyone does equally!

Is that a fair system?

Seriously, that sounds like Japan.

 

[Nithoniniel]

Gardiner's bad and the Corsi ghost is helping him. It's the only way people like him will accept the validity of advanced stats when they've already made up their mind about a player. An argument that will change his mind about Gardiner doesn't exist.

That's not true only on this topic though. People are notoriously bad at accepting that others might have a point, or that they could be wrong.

Another common thing, especially in this thread, is this:

These 'ridiculous' claims you're complaining about are claims conjured by you to create an argument nobody was, or wants to be, making.

Not only are some not open to possibly being wrong, they have no problem purposefully misinterpreting your posts if it can score them a point. Somehow that's what discussions on this board has become about, trying to "win". It's ridiculous, but so easy to get stuck in.

Thanks to you and Nithoniniel for helping me understand how to interpret these stats.

You're welcome! Happy to hear it, this is the reason I discuss all this after all.

 

[YOYOTCROSTER*]

Stats guys everywhere. Wow. Although Gardiner is the BEST skater on the team he is absolutely the WORST defensive player they have. It was proven again last night on the reverse 2 on 1. He flat out either does not know how to play it or he can't think fast enough to make the right defensive play. Either way that was a play every junior and minor pro player makes all the time. Maybe he was not taught defense as a kid or in the ncaa. Again either way it is real bad for the Leafs.

 

[Swayze*]

Stats guys everywhere. Wow. Although Gardiner is the BEST skater on the team he is absolutely the WORST defensive player they have. It was proven again last night on the reverse 2 on 1. He flat out either does not know how to play it or he can't think fast enough to make the right defensive play. Either way that was a play every junior and minor pro player makes all the time. Maybe he was not taught defense as a kid or in the ncaa. Again either way it is real bad for the Leafs.

He didn't start playing Defense until he was 17.

True story

Somehow he's the 2nd best shutdown guy in the league though.

Truth is he gets lost in his own end and doesn't have the IQ to figure it out

 

[Gallagbi]

Always makes me chuckle that the fans who talk most about hockey IQ seem to struggle with what it is and how to identify it.

 

[RLF]

He didn't start playing Defense until he was 17.

True story

Somehow he's the 2nd best shutdown guy in the league though.

Truth is he gets lost in his own end and doesn't have the IQ to figure it out

Sorry, got to ask. Who says he is the 2nd best shutdown d-man in the whole league? and if he doesn't have the IQ to figure it out...how did he become the 2nd best shutdown d-man in the league?

 

[Swayze*]

Sorry, got to ask. Who says he is the 2nd best shutdown d-man in the whole league? and if he doesn't have the IQ to figure it out...how did he become the 2nd best shutdown d-man in the league?

the advanced stats somehow show he's the 2nd best shutdown guy in the league.