Player Discussion: Winnipeg Jets Defense

[Whileee]

I am a fan of EW's work but I refuse to believe their Niku numbers. HockeyViz seems to align more with my eyetest and what I like to call my butt-clench test for d-men when they are on the ice.

An interesting point. While I like the concept of overall shot metrics in relation to performance and probabilities, I think they miss the "yikes!" factor that some D have.

 

[JetsFan815]

Looks like DeMelo is finally out of the doghouse. Took him half the season and horrendous play by Beau and Poolman to get there... maybe one day we will find out what he did around game 3-4 of the season to go from "Coach's porn" to so deep in the doghouse. Let's see what Morrissey looks like with him.

 

[boydkc]

Ya. It's really weird but from my analysis, watching, etc. Morrissey is the largest issue for the Jets.

Pionk looks like he can be the guy on a top4 pairing if he's given the right partner. I think he's meshed well with Forbort, which is a good sign as I think Samberg's top end is a much better Forbort.

So then you have one more top 4 pairing you need to figure out. Morrissey-DeMelo is fine to get into the POs in the ND or be a fringe PO team most seasons, especially with Jets strong Fs and Gs, but you need better there if you want to truly contend. DeMelo has looked to be the stronger of the two, although he's also the one that moves to softer mins when they are split, not the other way around.

I think DeMelo can be the guy for a pairing, but depends on the quality of the other top4 pairing and not sure that's the JEts. Forbort-Pionk, I don't think so. Samberg-Pionk, maybe but that's a lot of guess work.

Looking at both chemistry, microstats, and quality, I think Heinola actually projects as a solid partner for DeMelo.

If Heinola can be the stronger of the Heinola-DeMelo pairing, and Samberg an upgrade on Forbort, then you have a decent top 4 that can be placed in most situations: Heinola-DeMelo and Samberg-Pionk.

There's three issues with that:
1) You are dependent on the IFs with Samberg's and Heinola's development.
2) What do you do with Morrissey then?
3) There's nothing there in the short run as the Jets seem content letting Samberg and Heinola perculate in the AHL.

Maybe placing Morrissey on the 3rd pairing with everyone healthy will help him get back to his old self. I don't think Poolman or Stanley is the guy for him on the third pair, and definitely not Beaulieu. I think you need someone who likes to carry the puck more than Morrissey. He's fine at doing it but it seems he's done better when he's had more assertive guys like Trouba and Byfuglien, and even Myers to some degree.

Given how injuries work in the NHL, your 5th best guy plays top4 mins almost as much as they do on the 3rd pair. And Morrissey is a guy that can play PP and PK, but I think I'd lean Heinola and Pionk on the PP over Morrissey.

2021
Forbort,Pionk
Morrisey,Demelo
Stanley,Poolman
Beaulieu, Nogier
Niku,

2022

Morrisey, Demelo
Forbert, Pionk
Heinola, Gawanke
Samberg , Savard
Stanley,Poolman
Kovacevic,Chisholm

 

[garret9]

Just in case anyone is curious, AHL shots/gp

Heinola 1.88
Chisholm 1.56
Kovacevic 1.36
Gawanke 1.33
Samberg 1.0
Bourque 1.0
Sautner 1.0
Shaw 0.71
Oligny 0.69

 

[voyageur]

If Heinola draws in, where would you slot him in?

I'd probably go with Heinola-Pionk.
Pionk likes to have the play in front of him, and is ironically probably the most physical Jets d-man. Two puck movers with the right centreman could be good.

Morrissey-De Melo are a decent pairing, there's no offense from De Melo and Morrissey is a bit off offensively, but this is the closest the Jets could come to a true top 4 with an internal solution.

Forbort-Poolman

There's some size and Poolman can lead rushes, or join them. Forbort is integral to special teams, penalty killing. If Poolman wants to stay in the lineup he'll have to be good in this area.

 

[Guerzy]

If Morrissey has really regressed to a third pairing dman and does not bounce back, that will be such a shame. I'm just left wondering how this happened...

 

[Weezeric]

Depends on definition of top pairing guy.
My definition here isn't what I would normally use, so that's really my fault. I just mean a guy that can be the driver of the pairing in tough mins.

EDM: Nurse is having a good season results wise but he's been fairly stappled to McDavid so there may be some conflation there.
CGY: Is more by solid comittee than having top end.
MTL: Petry has evolved into that kind of guy, and Weber is still a fringe guy there.
OTT: Chabot has been good historically. This season is a bit of a gong show with them.
TOR: Rielley is all over the place and overrated but has his uses, and Holl and Muzzin can be complimentary piece to a top pair defensively. Ultimately they are flawed and not do-it-all types but still useful guys.
VAN: Scmidt I think is really good. Hughes has promise.
WPG: Pionk has been there this year so maybe he develops into a full on guy. I think part of it is chemistry with Forbort even if Forbort's overall quality level is low.

So the answer is... maybe.
I mean I'd put Pionk in top 10 for ND for defense, but it's definitely a division lacking in that position.

Would Ekholm-Pionk be the best pairing in the North division?

 

[blueandgoldguy]

Depends on definition of top pairing guy.
My definition here isn't what I would normally use, so that's really my fault. I just mean a guy that can be the driver of the pairing in tough mins.

EDM: Nurse is having a good season results wise but he's been fairly stappled to McDavid so there may be some conflation there.
CGY: Is more by solid comittee than having top end.
MTL: Petry has evolved into that kind of guy, and Weber is still a fringe guy there.
OTT: Chabot has been good historically. This season is a bit of a gong show with them.
TOR: Rielley is all over the place and overrated but has his uses, and Holl and Muzzin can be complimentary piece to a top pair defensively. Ultimately they are flawed and not do-it-all types but still useful guys.
VAN: Scmidt I think is really good. Hughes has promise.
WPG: Pionk has been there this year so maybe he develops into a full on guy. I think part of it is chemistry with Forbort even if Forbort's overall quality level is low.

So the answer is... maybe.
I mean I'd put Pionk in top 10 for ND for defense, but it's definitely a division lacking in that position.

Forgot Brodie in Toronto.

 

[Jet]

Update on model results...

EW RAPM xG and Corsi (in xG order):

Player	xG±/60	C±/60
Neal Pionk	0.199	2.78
Dylan DeMelo	0.073	0.36
Sami Niku	-0.039	-2.23
Logan Stanley	-0.046	0.31
Tucker Poolman	-0.093	-1.26
Derek Forbort	-0.132	-0.99
Nathan Beaulieu	-0.245	-3.44
Josh Morrissey	-0.251	-4.1

[TBODY] [/TBODY]

And HockeyViz Isolated Impact:

Player	Off	Def	Diff
Dylan DeMelo	6.3	-4.1	10.4
Neal Pionk	2.8	-1.6	4.4
Derek Forbort	-6.3	-3.9	-2.4
Logan Stanley	-1.2	2.9	-4.1
Tucker Poolman	-7.2	3.5	-10.7
Nathan Beaulieu	-6.7	4.4	-11.1
Josh Morrissey	-2.6	8.6	-11.2
Sami Niku	-8.5	10.6	-19.1

[TBODY] [/TBODY]

The later uses last year as a prior at the start of the season (although we are moving at the point where past season isn't having a huge impact anymore).


Some thoughts:

1) No doubt that DeMelo and Pionk are the two tredding above water. They are the two driving success from the back end. Jets need more help than that.

2) Forbort is what he is. Kills offense on both sides. There's utility for those types. It would be better if the difference was + and you probably don't want him as a top4 regular in the long run but he can be useful on a team with better top end than the Jets have.

3) Niku isn't struggling as bad as past years, but with 2 really bad years and this year just being sorta bad... it isn't much for promise.

4) Stanley has been sheltered and a negative, but he's a 1st year defender. Maybe he can be a fringe top4 guy, or maybe he'll just be a 3rd pairing guy. But, at least he doesn't look like the worst guy.

5) Poolman is not a top4 pairing guy.

6) Morrissey how much you've struggled. I know there's a lot of excuses out there for him, but if he's as good as many think he is he may struggle in particular usages but I still think he wouldn't be struggling as much as he currently is.

Jets still look like a team that has:
2-3 2nd pairing guys (Pionk, DeMelo, Morrissey IF he can bounce back to old him)
3 3rd pairing guys (Forbort, Stanley, Poolman)
2 pressboxers (Niku, Beaulieu)

Team needs a top pair.

It's funny you say that because I've thought that Forbort has been noticeable offensively the past few games. Not sure if that is a result of coaches getting him to adjust his game somewhat but I'm seeing him pinching down a lot more and saving plays. Then again, it seems our whole corps is doing that, so maybe he's just been more effective at it?

 

[Jet]

Looks like DeMelo is finally out of the doghouse. Took him half the season and horrendous play by Beau and Poolman to get there... maybe one day we will find out what he did around game 3-4 of the season to go from "Coach's porn" to so deep in the doghouse. Let's see what Morrissey looks like with him.

He wasn't in the doghouse. He was supporting a player with very little experience. Once Stanley earned some trust from the staff, they moved Demelo back to his usual spot, hopefully to prop up Morrissey.

 

[Adam da bomb]

He wasn't in the doghouse. He was supporting a player with very little experience. Once Stanley earned some trust from the staff, they moved Demelo back to his usual spot, hopefully to prop up Morrissey.

He was also there with beaulieu. Stanley still hasn’t trust despite excellent play. He played much less in 3rd.

 

[Jet]

He was also there with beaulieu. Stanley still hasn’t trust despite excellent play. He played much less in 3rd.

I think that has a lot more to do with the fact the Jets needed a goal.

 

[Adam da bomb]

I think that has a lot more to do with the fact the Jets needed a goal.

Even though he has more offensive ability than demelo and poolman

 

[garret9]

Would Ekholm-Pionk be the best pairing in the North division?

Yes

 

[garret9]

It's funny you say that because I've thought that Forbort has been noticeable offensively the past few games. Not sure if that is a result of coaches getting him to adjust his game somewhat but I'm seeing him pinching down a lot more and saving plays. Then again, it seems our whole corps is doing that, so maybe he's just been more effective at it?

Statistics are measures of a player's impact in the environment they are in, but that environment is also impacted by the decision they make. If Forbort is activating more because coaches have asked, it will impact results for sure. I think in general the Jets D will be better for it, but not sure on every case.

I think a good (albeit one of the most extreme) example is Cody Ceci. He either has good results or really, really, really bad. His results are good in seasons where he's not the primary puck carier and it's bad when he is.

 

[DRW204]

If Morrissey plays at a high-level i don't think we're in terrible shape. That's a big IF of course, given his last 2 years of play.

Forbot-Pionk has been a great pairing. Pionk has definite chemistry with Forbes and has flourished.

Id maybe get an upgrade at 3RD, a lower end acquisition that doesn't cost a 1st+, maybe Savard?

run

Morrissey-Demelo
Forbot-Pionk
Stanley-Savard

Heinola is the wild-card ofc if he can play RD at a good level.

i am just a bit wary on spending a 1st++ for Ekholm. and then having to re-sign him to a ufa deal at 32.

 

[garret9]

Update on results for 3 models:

Note:
HockeyViz uses past performance as a primer while Evolving Hockey and Top Down do not.
In other words, EH and TD (RAPM models) assume everyone is average at the start of the season and as the sample expands everyone moves out based on performance, while HV everyone starts between average and their previous season and as the sample expands they are moved out from that point.

There are pros and cons to each. HV likely better evaluates true talent for players when they disagree but not always, while the other two are more "what have you done for me." The RAPM models are more sensitive to changes (injuries, different usages, age) but are also more sensitive to "luck." They also struggle more when players are rarely separated (ex: Forbort-Pionk pair).

 

[Jet]

Even though he has more offensive ability than demelo and poolman

Does he?

 

[Adam da bomb]

Does he?

You are certainly entitled to your own opinion.

 

[Duke749]

Even though he has more offensive ability than demelo and poolman

We have no evidence as of now that suggests that. Getting shots towards the net isn’t some indication of above average offensive ability. Also, he’s not exactly the guy you want pinching in deep. It’s not his game in the slightest.

 

[Adam da bomb]

We have no evidence as of now that suggests that. Getting shots towards the net isn’t some indication of above average offensive ability. Also, he’s not exactly the guy you want pinching in deep. It’s not his game in the slightest.

And Demelo has above average offensive ability and blazing foot speed? Seen any beautiful pinches from Demelo?

 

[Duke749]

And Demelo has above average offensive ability and blazing foot speed? Seen any beautiful pinches from Demelo?

Downplaying DeMelo doesn’t suddenly make Stanley a viable option.

 

[Adam da bomb]

Downplaying DeMelo doesn’t suddenly make Stanley a viable option.

So there are no viable options other than Pionk , bort, and mo so play all the other guys and keep them fresher.
Also simply beautiful defence of Demelo just a masterpiece.

 

[Whileee]

Statistics are measures of a player's impact in the environment they are in, but that environment is also impacted by the decision they make. If Forbort is activating more because coaches have asked, it will impact results for sure. I think in general the Jets D will be better for it, but not sure on every case.

I think a good (albeit one of the most extreme) example is Cody Ceci. He either has good results or really, really, really bad. His results are good in seasons where he's not the primary puck carier and it's bad when he is.

Another consideration is that statistical measures of performance based on stochastic constructs (like expected goals) are sometimes hard to reconcile with rare events that are discrete (like goals), contributing to game outcomes that are also discrete. If games were determined by the xG winner, then it would be easier to link players' shot metrics to their contribution to a win or loss. But the reality is that a mistake by a player that leads to a goal can be much more consequential in terms of game outcome than is measured by shot metrics. As an example, the Jets had better shot metrics than the Habs a few games ago, but everyone knows that free pizzas from Copp and Pionk and Morrissey led directly to goals and the game outcome.

Now, some of those events are captured in expected goals measures, but they very often aren't. As an example, last night McDavid scored two goals on miscues by a Jets' D. According to Naturalstattrick, McDavid's total individual expected goals at 5v5 was 0.23. He scored 2 goals in a 2-1 game. That makes it seem that the Jets did a better job than they did, because we know he's lethal with a clear path to the net, and it's a big problem if your D allows it. So, how do we assess the performance of Forbort? His shot metrics were fine, and his aggressive pinching helped the Jets generate more zone time and shot attempts. But he also made a poor decision on a pinch and McDavid scored the winning goal.

I think this is one of the reasons that coaches and players sometimes are at odds with data analysts. They think of individual matchups and goals and ultimately game outcomes, which magnifies consequential events and diminishes a focus on statistics like shot attempt share and overall expected goals. Any coach would happily trade 0.23 expected goals in shot metrics for two clear breaks by McDavid.

 

[10Ducky10]

If Savard comes cheap enough...
JMo DD
Heinola Pionk
Forbort Savard
We would have Poolman, Stanley etc... for depth...switch Heinola/Forbort?