Klefbom is the most impactful d-man and a #1

[Ohashi_Jouzu*]

I'm not completely disregarding it, I'm just trying to get to the bottom of Klefbom scoring the highest in the league.

I do think he's an excellent player, but I think we'll all agree he's not the best player in the NHL.

Everyone has to remember that this simply measures what he did during the season, and doesn't necessarily speak to "potential", skills "ranking", or performance in a future season under inevitably different circumstances. Obviously he was REALLY effective this season. And if anyone, including him, routinely showed up near/at the top of any kind of composite "measure" like this you'd have to conclude that they are supremely effective players - or at least were over that period of time. But this is just one year, right? So you have to take for what it is without trying to say too much about the future (as is ALWAYS the case with anything stats related) until you have a pattern to look at.

 

[Aceboogie]

Everyone has to remember that this simply measures what he did during the season, and doesn't necessarily speak to "potential", skills "ranking", or performance in a future season under inevitably different circumstances. Obviously he was REALLY effective this season. And if anyone, including him, routinely showed up near/at the top of any kind of composite "measure" like this you'd have to conclude that they are supremely effective players - or at least were over that period of time. But this is just one year, right? So you have to take for what it is without trying to say too much about the future (as is ALWAYS the case with anything stats related) until you have a pattern to look at.

Klefbom was really high in a number of metrics even last year, albeit an injury shortened year

 

[Ohashi_Jouzu*]

Klefbom was really high in a number of metrics even last year, albeit an injury shortened year

True. Still small enough of a sample to justify anyone's skepticism to a degree, though, and he's obviously a player with much left to prove in terms of his numbers earning that kind of reputation.

 

[Spazkat]

On that list we see guys routinely ranked high by fans in terms for being good defensive D. Guys like Tanev, Brodin, Stralman, Vlasic, Weber, Doughty, OEL, Hamhuis. Do I think there is some exceptions even on this list? yes for example id question Lovejoy, Miller, Campbell. Also some guys probably missing. But more or less id say that list contains the right players. So for me Id probably say this list should highlight some players you probably didnt know were so good defensively, not use this a definitive ranking

It also contains a lot of guys whose partners are generally considered to be better than them - by a lot. Ellis (Josi), Stralman (Hedman), Manson (Lindholm), Klefbom (Larsson), Hjalmarsson(Keith), Ekholm (Subban). I don't think that's any coincidence.

 

[Soundwave]

All I can I say is thank goodness the Oilers didn't do what I wanted and taken Mark McNeill at that spot

Got very lucky.

 

[kurtcobang]

All I can I say is thank goodness the Oilers didn't do what I wanted and taken Mark McNeill at that spot

Got very lucky.

I wanted McNeill too and just to one up you I was begging for Bennett over Drai despite the fact Bennett had the upper body strength of a child
I was all in on the future Dougie Gilmore and I was disappointed when they called Drai's name lol

 

[NickLidstrom]

Klefbom would be on the top pairing in Calgary and Larsson would be the teams best shutdown defenceman.

Brodie and Gio are both better.

 

[Filthy Dangles]

so what goes into the 'Defense' component of GAR? And, for that matter, what goes into all the components that comprise it?

Is it just shot suppression? How does it weigh Team effects? Like what if you play on a team that concedes a lot of shots in general but you are good at suppressing relatively(Relative to your other teammates)? Do you get more credit than a good shot suppressor on a good team because of the 'relative' boost?

Is Shot Quality and Type measured? Maybe you give up some shots but they are not high chances or quality. I've read that this is a repeatable skill IIRC, that is skaters can impact shot quality. Maybe you block a lot of the shots that your team gives up when youre out there. Do you get credit for that?

Where's the documentation on how exactly one arrives at these numbers? Seems like a big black box that just spits these out and we are just forced to run with them.

 

[bucks_oil]

Not 100% on this but basically on the PK players have very little agency, the goalie has the most and only impact, the skaters' impacts in front of him are negligible.

As a goalie myself.... complete hogwash.

 

[bucks_oil]

My mistake. In the 15-20 or so times I've watched him play, I've never seen him uncork one like that, so I figured he had more of a muffin shot as the ones I did see him take were just okay shots, nothing like the one in the YT vid above.

In fairness he is very selective about when he uses it. Usually when he shoots, he's just trying to get a wrister through. Not sure if this is a function of Oilers pp or his decision making, but I'd sure like to see more of his shot unleashed.

 

[THall4]

OP losing his mind with some of his posts.

Klefbom still has his inconsistencies, esp on the defensive side...I'd give him a few years before I call him a true #1. But his play in his shortened year in 2015-16 and his play of 2016-17 certainly suggests he's trending that way.

 

[Aceboogie]

so what goes into the 'Defense' component of GAR? And, for that matter, what goes into all the components that comprise it?

Is it just shot suppression? How does it weigh Team effects? Like what if you play on a team that concedes a lot of shots in general but you are good at suppressing relatively(Relative to your other teammates)? Do you get more credit than a good shot suppressor on a good team because of the 'relative' boost?

Is Shot Quality and Type measured? Maybe you give up some shots but they are not high chances or quality. I've read that this is a repeatable skill IIRC, that is skaters can impact shot quality. Maybe you block a lot of the shots that your team gives up when youre out there. Do you get credit for that?

Where's the documentation on how exactly one arrives at these numbers? Seems like a big black box that just spits these out and we are just forced to run with them.

XPM which is adjust plus minus. Detailed analysis by the creator here: https://hockey-graphs.com/2016/10/25/expected-plus-minus/

Response Variable
Rate at which an Expected Goals event took place
Explanatory Variables
All players on the ice
The coaches of both teams
Zone the shift started in
A score adjustment
If one of the teams is playing a back-to-back
A home ice advantage adjustment

The expected goal part of the variable takes into consideration shot distance, shot angle, shot type, shots of rush and on/off wing

[Expected Plus-Minus] does NOT measure how well a player has performed this season.

Statistics can be divided into two categories. Descriptive statistics tell us about what happened in the past. For instance, I can check how many page views this blog post has. Predictive statistics try to forecast what will happen in the future. I could create a model that estimates how many page views I’ll get over the next 24 hours. This difference between these is subtle, but important.

[Expected Plus-Minus] is meant to be predictive. It’s interested in how well a player will perform in the future, rather than what he did in the past. [XPM’s] emphasis on prediction explains why it uses some of the tricks it does.

For instance, I mentioned earlier that [XPM] uses data from [the previous season] in its priors. If my primary goal is to evaluate how well a player did this season, it wouldn’t make a lot of sense to use data from [last season]. However, if I want to predict what will happen in the future, the older numbers can help differentiate between players who have been consistently good (and will likely keep being good) and players who are merely going through a hot streak (and will likely regress to their mean).

GAR is predicative and below is the year to year repeatability of it. For me, I value predictive stats a ton because I want to know what to expect in future

GAR is not entirely black box as it uses stats scraped from NHL.com. So its not using privately sourced data like actual black box stats use. The only thing really black box is how those stats are configured but I do beleive if you were a statistican you could replicated the regression model and get to same stats

 

[wintersej]

When McDavid and Kelfbom are on the ice together, the Oiler score over 4 goals at 5 on 5 per 60. Without McDavid, but with Kelfbom, the Oilers are around 2 goals at 5 on 5 per 60.

McDavid skews Kelfbom's stats significantly.

 

[83to48]

When McDavid and Kelfbom are on the ice together, the Oiler score over 4 goals at 5 on 5 per 60. Without McDavid, but with Kelfbom, the Oilers are around 2 goals at 5 on 5 per 60.

McDavid skews Kelfbom's stats significantly.

Or they benefit from each other considering Klefbom is our best puck mover?

 

[Aceboogie]

When McDavid and Kelfbom are on the ice together, the Oiler score over 4 goals at 5 on 5 per 60. Without McDavid, but with Kelfbom, the Oilers are around 2 goals at 5 on 5 per 60.

McDavid skews Kelfbom's stats significantly.

And Mcdavid drops to 3.19 GF/60. For nearly every other D he does better away

McDavids GF/60 is the highest with Klefbom/Larsson and drops for every other D pair and forward. As much as McDavid helped Klefbom, Klefbom helped McDavid as well

For the record, 4.18 GF/60 (Klefbom and McDavids GF together) is absurd. The highest GF/60 in the league is 3.87 (for regular players)

 

[Seanaconda]

I wanted McNeill too and just to one up you I was begging for Bennett over Drai despite the fact Bennett had the upper body strength of a child
I was all in on the future Dougie Gilmore and I was disappointed when they called Drai's name lol

To be fair to Bennett he had a shoulder injury he was hiding when he couldn't do a chin up.

He missed like the entire year after that because of it but that's all he will be remembered for

 

[Pinkfloyd]

If we're going with the context that most impactful defenseman doesn't mean the best d-man in the league then we're essentially saying that whoever is the highest event defenseman in the league is the most impactful d-man. I would think that would be Burns. That impact isn't always positive but it's probably the most. lol

 

[lomiller1]

so what goes into the 'Defense' component of GAR? And, for that matter, what goes into all the components that comprise it?

The two major components to the 5v5 numbers are XPM and BPM. XMP is shot metric type data, BPM is box score type data. For the defence score of D-men it’s weighted 100% XPM since the BPM aspects do not predict anything.

Is it just shot suppression?

It’s not just shot suppression, it uses the only really good shot quality metric available. XPM is based on expected goals (@DTMAboutHeart’s version, which is the only one shown to outperform plain old shot attempts). A statistical model is used to figure out how much of each expected goal

How does it weigh Team effects? Like what if you play on a team that concedes a lot of shots in general but you are good at suppressing relatively(Relative to your other teammates)? Do you get more credit than a good shot suppressor on a good team because of the 'relative' boost?

It looks at all 10 players on the ice, all the players who they have been on the ice for a shot attempt, and so on. Basically it uses every shot attempt in the entire league all season as context to attribute a share of each shot attempt to each player on the ice. This is where it becomes a bit of a black box. You can’t really replicate this on the back of an envelope.

s
Is Shot Quality and Type measured? Maybe you give up some shots but they are not high chances or quality. I've read that this is a repeatable skill IIRC, that is skaters can impact shot quality..

As I mentioned above @DTMAboutHeart is really the only shot quality metric that has been shown to add value beyond what you can get from just looking at shot attempts. One of the main things that allows it to do this is that it considers shooter history. Eg, on shots of similar type & location Stamkos scores about 40% more often than the typical NHL player. This is the highest “finishing factor†of any player in the NHL. The expected goals model being used here factors this in, in addition to more typical things that separate rebound chances, rush chances tips, etc.

s
Where's the documentation on how exactly one arrives at these numbers? Seems like a big black box that just spits these out and we are just forced to run with them.

https://hockey-graphs.com/2016/10/24/a-primer-on-dtmabouthearts-war-model/

 

[wintersej]

Or they benefit from each other considering Klefbom is our best puck mover?

And Mcdavid drops to 3.19 GF/60. For nearly every other D he does better away

McDavids GF/60 is the highest with Klefbom/Larsson and drops for every other D pair and forward. As much as McDavid helped Klefbom, Klefbom helped McDavid as well

For the record, 4.18 GF/60 (Klefbom and McDavids GF together) is absurd. The highest GF/60 in the league is 3.87 (for regular players)

Don't deny at all that there is a big dropoff in the EDM d-group when Kelfbom and Larsson are not on the ice. Just pointing out that Kelfbom's overall "offensive impact" numbers compared to other D around the league get a real bump playing with one of the top 2 offensive players in hockey. I think, if you have a McDavid, you can have Kelfbom as your #1 and be a contender... but he is a mediocre at best #1 and as much as GAR is great great work, I don't think it takes into account the McDavid effect nearly enough.

 

[Panda Bear]

How does Klefbom look without Larsson this season and vice versa?

 

[Aceboogie]

Don't deny at all that there is a big dropoff in the EDM d-group when Kelfbom and Larsson are not on the ice. Just pointing out that Kelfbom's overall "offensive impact" numbers compared to other D around the league get a real bump playing with one of the top 2 offensive players in hockey. I think, if you have a McDavid, you can have Kelfbom as your #1 and be a contender... but he is a mediocre at best #1 and as much as GAR is great great work, I don't think it takes into account the McDavid effect nearly enough.

Fair enough. His biggest GAR aspect is defensive actually tho, ranks #1 for all D. Not saying he's the beat defensively in nhl but super underrated

 

[Freudian]

Sounds like Oiler fans have soured a bit on Klefbom lately. Of course that seems to be using the good old eye test, which we all know is useless.

 

[ThirdManIn]

Doughty isn't the same as the mini-Kings dynasty Doughty anymore.

Burns' even strength defense is negatively impacting his team.

Karlsson doesn't carry his defence corps the way Klef carries his.

I like your argument style. Make claims so bold they overlap into the absurd, and provide no actual evidence to support those claims. Even in a time after the death of expertise, that is impressive.

 

[Bjornar Moxnes]

Sounds like Oiler fans have soured a bit on Klefbom lately. Of course that seems to be using the good old eye test, which we all know is useless.

Actually advanced stats speaking Klefbom has also been bad which wasn't the case when Oilers fans were overrating him, he was legitimately great during that time period.

 

[zeke]

Never say a player is frontline quality until he proves he is in frontline usage.

klefbom looked great in secondary usage last year, but isn't handling legit frontline usage well at all this year.