Jets Forward Usage: A Deep Divide

Just as Ryan Blight’s scathing piece about Kevin Cheveldayoff’s lack of changes to the Atlanta Thrashers came to define how we thought about the Jets in the summer, Tim Bonner’s recent piece regarding the massive gap in talent between the team’s top group and depth players is coming to define how we talk about the next stages of this team. 

The conclusion of Bonner’s work is the culmination of conversations he, we, and every Jets fan has had over the last 10 weeks. This team has problems, but the problems are contained to specific areas, namely the bottom-6 forwards, and bottom 2 or 3 defencemen.

One challenge this team faces is that it has to fix a number of concurrent deficits in order to even diagnose whether it has larger issues with core players like Evander Kane, Ondrej Pavelec, or Dustin Byfuglien.

A contention I stand by at this stage in the season is that the first change has to be Claude Noel. Inside, we’ll see just how his lack of line matching and line awareness have impacted the performance of the forwards, with some extra sensitivity to Bonner’s point about the deep divide between capable and incapable players on this roster.

The Chart 

The chart is easy to read once you know it. All the stats use relative corsi, so we are comparing how the Jets do compared to each other. Each bubble is a player. Blue bubbles are positive corsi (shot attempts for – shot attempts against), orange are negative corsi. The size of the bubble is how big that number is (so Halischuk is a big minus player, Thorburn only a slightly minus player). The closer to the top the bubble is, the harder the competition they face on average. The closer to the right edge it is, the more of their shifts start in the offensive zone. 

Instant Analysis

I used a 10GP cut-off, because the extras take away our focus from the core group. Jim Slater is the most important player excluded, but also John Albert, Patrice Cormier, and now Eric O’Dell have all been left off as well. You can see it unedited here from Ninja Greg.

The pattern is extremely obvious – a cluster at the top of the chart contains all top-9 forwards. The trio at the bottom face almost exclusively 4th line opponents, and Thorburn has been pulled into no-man’s land by his dual assignment. We know Frolik played a tough assignment with Jokinen to start the year, while Scheifele got the easier minutes. Slowly, those have evened out to the lines we know today, where Frolik is the mentor to young Scheifele on the third line. Meanwhile, we know Blake Wheeler got a push from the coach to score when he was struggling early, and he leads the top-9 pack in zone starts. As well, moving up and down the roster as he does, his competition has been a smidge less intense than Ladd and Little.

The clustering tells us that Noel is not matching his lines. If he was, different lines (or even just different players) would be spread out vertically on the chart. You can look at Chicago’s for an example of what I mean. We know acedotally that line matching is a requirement in the exceptionally competitive Western Conference, and I’ve long argued that Noel’s inability to do it is why the team struggles against shallow groups like Dallas.

As well, most coaches uses zone starts to give their offensive guys a push (as Alain Vigneault became famous for with the Sedins). Noel isn’t controlling it much (again, we see the clustering), but is using offensive zone faceoffs to shelter Anthony Peluso, as well as Halischuk and the offense impaired Tangradi. Travis Hrubeniuk has talked on this blog about how the team generates as many shots from offensive zone faceoffs as defensive zone faceoffs. This is one of the many reasons why. 


A Deep Divide

With regard to ability, what we see above is exactly what we know to be true by eye. Scheifele is in over his head (until very recently), Frolik is put in a position to struggle, while all of Halischuk, Wright, Peluso, Slater (not shown), and to a lesser extent Thorburn are ineffective at best.


As Armchair Coaches, we can identify three main problems with this usage:

  1. The Jets’ third line was drowning against this level of competition. Matt Halischuk, in particular, was not doing well before his injury. He’s giving up a worse shot attempt differential than every player but pressbox warmer Anthony Peluso. 
  2. Devin Setoguchi is a very successful tough minutes players and is one of the coach’s whipping boys this season. We already knew Noel didn’t monitor corsi numbers (Burmistrov taught us that lesson well), but given Setoguchi’s offence, this is even more problematic. Cheveldayoff has said that Frolik and Setoguchi were targeted additions in part because of their experience in the West and with the Central division. Frolik has been in an uphill battle for ice time all year, and Setoguchi is getting benched despite being the most successful Jet against the hardest opponents. What more can he do?
  3. In fact, Seto is third in rel corsi to only Evander Kane and Eric Tangradi. Stop me if you notice a pattern in those three. Oh, they all face the scorn of their hapless coach, you say? More than that, they reveal that this coach isn’t actively trying to manage the puck or control territory, and he’s not aware of who can do what jobs well. It’s all instinct and waiting for ‘A’ level effort.

I think the solution is obvious even without changes to the roster. But whatever your solution, the problem is self-made by this coach, and he stuck with it for far too long. 

Defender Effects 

The following table comes from the work of Tyler Dellow, a well known Oilers blogger and now writer for Sportsnet. He’s long been critical of Ondrej Pavelec and has done sporadic work around the Jets in the past. The table is his attempt to quanitfy Bonner’s point about the divide in quality. It matches forwards (vertical) with defenders (horizontal) to present their combined corsi score when those players are on the ice together. The red cells are above 50% corsi share (more shot attempts for than against) and the blue is below 50%. 

Dellow instructs us to notice how things turn red to blue after the top six forwards, but it’s worth noting that Stuart is all blue, and that Ellerby and Trouba are almost as bad. Trouba’s excuse is that he’s a rookie. Ellerby’s excuse might be that he plays with Mark Stuart. Mark Stuart’s excuse is that he already tries so hard.

Putting aside Bogosian’s small sample of minutes for now, the problem is in two dimensions. The bottom-six forwards drag down the defence (all of Buff, Enstrom, Clitsome do worse), and the bottom three defenders drag down the forwards. 

The Summative Consequences

It’s clear at this point that the Jets have limited ability in their bottom six forward group. It’s not just a lack of raw scoring skill, it’s the fact that a rookie centre is expected to play defensive minutes, and Nashville’s castaway Matt Halischuk is slotted as a defensive player with no evidence he’s ever done such a thing in his career. It’s that James Wright, Jim Slater, and Anthony Peluso are extoled for their defensive acumen, but spend all their time in their own zone against the dregs of the NHL. 

It’s not an ideal cast of characters, but the coach is deploying them in a sub-optimal arrangement, and we’re left with the question of how much a better coach could get out of this group. 

  • Kevin McCartney

    Great analysis! The advanced stats stuff has grown on me over the last little while – that table at the bottom perfectly explains what we’ve all been seeing on the ice. Hopefully the Evander Kane haters take notice. Also, supports what I commented a week or two ago – Scheifele either needs to be protected more, or (preferably) learn the pro game in the minors.

  • Kevin McCartney

    Good article. The things that stood out the most for me were: (1) How outmatched the third line has been (with the exception of the past two games, which we sure hope is the start of a trend in the desired direction), (2) Setoguchi’s relative success, despite being one of coach’s black sheep; and (3) Olli’s relative success this season, despite his high percentage of defensive zone starts. I also think your point is a fair one that we really need to fix some things before truly being able to evaluate whether there are problems with our top players.

    I’ll admit it surprised me to see that Scheifele has had as much of an offensive zone push as the chart indicates. I thought we’d been seeing the opposite (maybe based on how much sympathy I’ve felt for him, time and time again taking face offs in the D zone). But if that’s the case, why do you think the chart showing such a discrepancy between Scheifele and Frolik in where they’re being started?

    • Kevin McCartney

      If, by the Grace of God, KSF stays together for the duration I would love to see some 35/35 split stats at game 70 for this team. Scheifele is now in the perfect position to succeed with a combination of his improving comfort on the ice and being put on a line with two possession masters of the universe in Kane and Frolik (and with a little luck, favourable deployment).

      • Kevin McCartney

        Agreed. For about 36 seconds in the last game, Scheifele gave me reason to feel genuinely excited watching this team. I haven’t felt that way about the Winnipeg Jets since……maybe since the old Alexander Burmistrov days. If that’s not too nostalgic.

    • Kevin McCartney

      Thanks for the comment, Dan.

      I think the split in zone starts from the third line is just from previous line groupings. Frolik was with Jokinen in that tough minutes pairing at the start of the year, while Scheifele was with Kane and Seto. Meanwhile, Halischuk spent time on the 4th unit with Wright and Slater/etc. They got a push out of protection, I think.

      X had that idea about the 35 game splits, and I’ll try that. I’ll also look for a way to parse it out by lines over the season as best I can, so we could see Frolik with Jokinen vs Frolik with Scheifele.

  • Kevin McCartney

    I am only half way through, but before I lose my train of thought (short train on narrow track, let me tell ya).

    “The conclusion of Bonner’s work is the culmination of conversations he, we, and every Jets fan has had over the last 10 weeks. This team has problems, but the problems are contained to specific areas, namely the bottom-6 forwards, and bottom 2 or 3 defencemen.”

    That, my dear author, is a high estimation of what “every” Jets fan has for conversations. Outside of the cosy confines of this outpost and AIH the conversation is mostly about how terrible Buff is and just how much more face-punching is required to win.

    Devin Setoguchi has actually been something of a pleasant surprise to me, as a career negative Corsi player he has produced reasonable possession numbers and has looked fairly good to the eye. The Thorburn experiment I would expect to be torture on his numbers and his patience. This is the difference I suppose between a negative Corsi player and a Corsi sinkhole like Halischuk.

    I am not sure it is fair to say that Claude Noel does not match lines, I would say that it is fair to say that he believes zone starts are more important, that is for certain. What Noel does seem to do, to my meagre observations that are infrequent and usually interrupted by child-rearing and household duties, is match in a manner that is fluid throughout a game and from game to game. He appears to have the belief that particular players or lines are “going” on particular nights and thus match on that basis and change those matchups/deployments as the game progresses sometimes. The net result is a chart that looks like randomness but is a product of a game-by-game selection. (The only really constant being that he has the good sense to not trust his forth line at all, so it is easy competition and Oz start, credit where credit is due.) One observes the great outcry early in the season when Scheifele was “benched” in the 3rd a few times when what really seemed to happen was the Jets were playing so badly they had no Oz starts to give him and Noel didn’t want to play him otherwise (again, credit where it is due, I think that was the right move at the time, the kid was pretty mediocre then).

    I am too lazy to look it up but is seems to me that the deployment/QualComp charts for Noel in 2011 were quite different than this, it is possible that we are seeing more flailing about on his part or that this is a deliberate strategy of some kind. (Perhaps Thorburn on the 2nd line was some kind of stealth tanking subterfuge that the players are slightly too talented to let work to its fullest?)

    This is where I stand with Claude Noel, his ideas about players being good one night and bad the next are absurd, thus we observe on the ice a by-product of this absurdity. (I have no idea what sort of a player development coach he is but I have no reason to believe he is a bad one.)

    • Kevin McCartney

      haha True point about who is talking about what. To be honest, I try to listen to sports radio sometimes and just can’t do it. From the ‘James Wright is a good 7th man’ comment in the pre-season to the talk of Thorburn literally being better than Evander Kane a week ago, it’s just wretched.

      So in my brain, that narrative of ‘pucks optional’ is heavily marginalized. I’m saddened to hear it’s not marginalized in the real world.

      It’s funny, actually. The Jets have had a top-5 team for size every year. But it’s Burmistrov, Wellwood, Setoguchi, and Frolik that have made this team work over the years. They’re about one face puncher away from drafting Aaron Ekblad, not from making the playoffs.

      Anyway, more interestingly, I’m really intrigued by your point about Noel’s line matching pattern. He does favour some lines on some nights, but I’ve never noticed a pattern where he controls their matchup. I always thought (maybe assumed) he was just putting them over the boards more and letting the other coach pick his match. I haven’t designed a data test for this yet, but I’ll work on it. Maybe rolling game relative Qual comp? I think that could be a fascinating article. Thanks, X.

      • Kevin McCartney

        “I haven’t designed a data test for this yet, but I’ll work on it. Maybe rolling game relative Qual comp? I think that could be a fascinating article.”

        That would be an interesting article.

        If @X is correct, then it really isn’t line matching at all, is it? I’m no expert, but I would guess it would be tough to play in a system such as that – game planning, understanding/executing on your assignment or just getting/staying in the flow of the game night-to-night would constantly be in flux, no? Doesn’t sound like a recipe for success.

  • Kevin McCartney

    Re: Dallas Example

    I thought it was funny you noted Dallas in particular because they picked up Ruff, who actually *prides* himself in randomly throwing lines on the ice.

    (I have no idea how this link is going to render, forgive me.)

    I wish the Ninja offered a fixed scaling option, then one could do a decent job of comparing them without frying your Hippocampal place cells en-mass.