Just as coaches focus on who a player’s teammates should be, they also focus on who they are matched up against in a game. It is one thing to put two defensemen together; it is another to decide whether they should play the most minutes and/or take on the other team’s best players when possible. While its impact may be overblown, who a defenseman plays against matters to a degree. In this post, I will explore how the New Jersey Devils defensemen performed against their competition.
One of the more interesting methods to determine competition has been the WoodMoney metric for competition. It is named this because it is the work of two Oilers fans who go by Woodguy55 and G-Money. Woodguy55 has a full explanation of how it was devised and how it works at his blog, Because Oilers. The basic idea is that players are divided up into three categories. The top players are “elite,” most everyone is in the “middle,” and the lesser players (think fringe players, fourth liners, etc.) are in the “gritensity” level. A player’s on-ice Corsi stats (plus relative stats) are divided up by how they did against each of the levels of opponents. While binning tends to create grey areas about whether players are assigned appropriately, the method makes intuitive sense. If I want to know how a defender played against the better players in the league, the WoodMoney method makes that clear.
If this seems familiar to you, then you have a pretty good memory of this site. I’ve used this metric before in 2016, back when the data was distributed through a comma-delimited spreadsheet. Recently, their stat site PuckIQ was re-launched with some new people helping, new names for the competition levels (gritensity is better sounding the dregs) and a much slicker front end. Now is a good time as any to take a look at how the Devils’ defensemen stack up with this methodology for quality competition. Let’s find out another means to praise Will Butcher, a reason to give a little more credit to Sami Vatanen, and why Connor Carrick is a better option today than Steven Santini among other findings.
All data is from PuckIQ as are their abbreviations as defined in their glossary. I limited the data to defensemen who have played over 200 minutes of ice time. So Eric Gryba, Colton White, and Josh Jacobs missed the cut. Sorry to those three.
2018-19 Devils Defensemen Time on Ice by Competition Level
Let’s begin by establishing how often each defenseman saw each of the three levels of competition:
Surprising nobody, Andy Greene and Damon Severson saw a plurality of their ice time against the highest of the three levels. The two played a lot together last season and so they would have similar proportions of competition. Greene moreso than Severson, which shows how the coaching staff still leans on #6 for better or for worse. What is a small surprise is how often Sami Vatanen saw the elite level of competition. As per Natural Stat Trick, Vatanen did not play a lot with Greene last season so it is not a case of anyone riding with Greene saw a bump in Elite TOI%. It may be more of a function of who opponents target than anything the coaches are doing.
The rest of the defensemen have the middle level as their most common level of competition. This makes sense. Greene-Severson was utilized as a first pairing that often saw the opponent’s better players. Everyone else would mostly see, well, everyone else. Even the third pairing defensemen like Connor Carrick, Egor Yakovlev, and Steven Santini saw more of that middle level than the lowest level of “gritensity” levels. What I found interesting is that the coaches still sought to protect Will Butcher to a degree. He nearly saw as much ice time against that lower level of competition as the middle level. Only Santini had a higher proportion of shifts against the “gritensity” guys and that was just by a little bit.
2018-19 Devils Defensemen’s Corsi For Percentage, Corsi Against per 60, and Relative Corsi to Teammates by Competition Level
Let’s begin with CF%. This stands for Corsi For percentage. The percentage represents the proportion of the Devils’ shot attempts over all shot attempts when that player is on the ice. As a whole, the Devils were one of the worst teams in the NHL (29th, to be exact) in CF% as a team according to Natural Stat Trick. Most Devils finished well below 50% in CF% against all competition levels. Would it be any better with a breakdown by competition? Here is how each of Devils defensemen performed in CF% by competition level:
To answer the previous question, no, not really. When any of the Devils defensemen were on the ice against elite competition, the Devils were out-attempted. Only Will Butcher and Mirco Mueller topped 47% CF%, which is still really low but they were better than the rest. Knowing that Vatanen played quite a bit against elite competition, his 46.4% CF% against them suggests that he was a better option than the other two defensemen heavily used against that level of competition. Greene’s 43.1% CF% against elite level of competition further suggests that the Devils suffered with his usage. Severson’s 44.3% was not good but still a step ahead against Greene. This makes me want to consider whether the first pairing should be Greene’s Replacement-Vatanen instead of Greene’s Replacement-Severson. Seeing how Mueller and Butcher did better in smaller sample sizes, it might be worth a try after how last season went.
The WoodMoney method reflects well for Vatanen and Butcher last season in that they both had a CF% of at least 50% against both middle and gritensity levels. Other defensemen cannot claim the same. From a CF% perspective, Ben Lovejoy did quite well against the middle level of competition; but struggled against the other two - including the weakest level. The team’s CF% was the highest when Mueller got to play against the lowest level of competition, but that took a step back against higher levels. Yakovlev had a breakdown like Butcher where his CF% against the middle level was better than his other two except the team was nowhere near breakeven in any of those levels. Speaking of defensemen who did not breakeven at any level of CF%, a chart like this shows that Carrick should be kept to playing against the lowest level and that has been far, far better in all three levels than Santini. It supports the idea that between the two third-pairing caliber blueliners, Carrick may be the better defenseman.
Since we’re discussing defensemen, we can see their CA/60 or attempts against the team per 60 minute rate. Not everyone plays the same amount of minutes, so a rate stat is more appropriate to compare with other players. CA/60 may not be the very best way to determine whether a defenseman has been doing their job, but it is a way to get an idea of how many attempts the team faces when that player is on the ice.
This chart clearly demonstrates how good the elite are. Only the limited minutes of Yakovlev saw the elite not generate as many attempts against the Devils as when Yakovlev was on the ice. That opposition’s lower level generated more attempts is both odd to see and perhaps backs up why Yakovlev could not secure a regular spot in New Jersey. The expectation is that the third-pairing defenseman can handle his level of competition. With a Gritensity CF% of 42.4% and a CA/60 higher than the other two levels, it does not appear that expectation was met.
I was pleasantly surprised to see Mueller have a CA/60 less than 60 and showed even lower rates against lesser competition. I would not go as far as to say that Mueller should be in Greene’s spot. However, he is far from a liability. On the flipside of Mueller, the Devils were caved in the run of play when either Lovejoy or Santini was on the ice. Not that the Devils were a stingy outfit, but their against-rate stats for attempts was not outlandishly poor. But it was when those two were on the ice. It informs why they were kept to easier competition, which worked quite well for Lovejoy in this regard. It helped Santini too and these rates are better than Carrick’s. Which means that Santini’s low percentages are due to the reality that the Devils’ offense dries up when he is on the ice. That may warrant a closer look on its own. Speaking of comparisons, while their rates are not elite, Severson again outperformed Greene in this regard.
Also once again, I look at this chart and I’m smiling at the numbers for Butcher and Vatanen. While we know that Butcher played more against the middle and lowest levels, this shows that the team was not giving up a high rate of attempts against. Especially against that lowest level. It is a sign that Butcher’s performances are above that level as he crushed it. As for Vatanen, he had a better CA/60 against elite competition than Greene and Severson, which is another point in favor of giving him tougher assignments. He also took care of business against the other two levels. While not to the great degree of Butcher, it is encouraging that he did this despite often being behind the 6-28 pairing.
Now, CF% against a competition level on its own tells us something about how the team performs against that level of competition when the player is on the ice. There is also a CF%RC stat, which is for the relative CF% of the player to his teammates for that level of competition. In other words, when Andy Greene steps on the ice to play against elite competition, how does the CF% for elite competition change? This, and more, can be revealed in this chart:
Look at Will Butcher and Sami Vatanen. That both are positive across the board is a good thing. When they stepped on the ice against any level of competition, the team’s CF% improved. The one to be really happy for is Butcher. The team’s CF% made some good gains when Butcher stepped onto the ice no matter what. A +1.5% difference is not nothing over the course of a season. It is a lot more valuable to a team’s performance, especially on a team that struggled in CF% as a whole all season long. The Devils need more performances like Butcher’s and, to a little lesser extent, Vatanen’s.
As with the other charts, Ben Lovejoy and Mirco Mueller should get a little credit for representing some good gains for various levels. Mueller having a positive relative CF% for elite level competition is intriguing. I do not know if it is enough to warrant a bigger role given how similar gains were not made against lesser competition. But in a pinch, he was not an albatross against the other team’s best. Lovejoy - likely with Butcher - also represented a very positive gain against mid-level competition. It further supports how he did fairly well in his limited role; although there were some limitations as seen with the -2.8% CF% change against the lowest level of competition. Egor Yakovlev had a Lovejoy-ian like split across levels. Except Yakovlev stands alone with that terrible -9.8% CF% change against the weakest level of competition. Given that he did not play a lot when he was dressed, that did not help the Devils very much. No wonder he did not stick around. But, hey, they had notable gains.
In between Connor Carrick and Steve Santini, this chart provides more evidence the Devils should keep both of them away from elite level competition as much as they can. It also shows how Carrick has been more effective. Against the other two levels of competition, Carrick did not yield a big swing in CF% in either direction whereas CF% dropped like a stone when Santini was on the ice against any competition. Again, this may warrants a more detailed post, but when Santini was on the ice for the Devils, New Jersey’s offense wilted to a point where the team’s CF% (and other such stats) plummeted. This is not the only means to judge a defenseman, but I am becoming more and more convinced that if the decision has to be made internally for a third-pairing right-side defenseman, then Carrick should be the one to get the spot over Santini.
As for the first pairing, this chart shows that Greene and Severson have similar-looking results across all three levels. However, Severson’s relative CF% values are better than Greene’s. Not that they are good, but it is better than the captain at each level. Especially at the elite level; whereas Greene coming on the ice led to a notable drop in the team’s CF%, Severson’s impact was just slightly below breakeven. It furthers my thinking that Greene may have held Severson back to a degree last season. Until the Devils find a replacement for Greene’s role on the team, this may continue. Good luck carrying the pairing, Severson.
Final Thoughts & Your Take
As a whole, the Devils really needed more play drivers among their defensemen. Butcher and Vatanen were the best the Devils had last season, and Vatanen missed quite a bit of time last season. While it is not everything, that their CA/60 at each level was not nearly as jarring as other defensemen shows that they may have had a good defensive impact. Still, they were not enough to offset the constant pinning back that Santini, Carrick, Greene, Severson, and Mueller (to a slightly lesser extent) faced. And against elite level competition, the Devils were out-attempted (and usually out-played) across the board.
Good defensive play is more than just keeping the opposition to the outside or making an important stop on the opposition. The defenseman needs to be able to make zone exits that facilitate offense, execute passes, and support the offense by more than just standing at the point to keep pucks in a play and fire low-danger fifty-five footers from the point. CF% is not a perfect measurement by any means but it points to more than just what is allowed and the Devils needed more of it from their blueline in 2018-19. If you believe the Devils defense is a problem, then this supports you.
As an aside, this is where analytics kind of butt heads with each other. The ones that focus on shots, attempts, and possession point to the Devils defensemen not being good. The ones that are based on shot location like expected goals or regression models driven by expected goals point to the Devils defense being actually good or at least decent. I wrote about this back in February and the following two months did not make huge changes in terms of what kind of shots they have allowed. This came to mind while finishing this post and I figured it would be worth mentioning.
Getting back to the WoodMoney method of competition levels for the 2018-19 Devils defensemen, I also think the Devils may need to re-think their usage some more. John Hynes kept leaning on Greene and Severson to take on elite level competition, with not-nearly-elite results. However, Vatanen having over 40% of ice time be against elite level competition is a sign that they are willing to try. They should continue it. Ideally, someone else has to take Greene’s minutes so Greene can hopefully play better against lesser competition - similar to what we saw with Lovejoy over the last three season. Or the coaches stop having a quasi-designated “tough minutes” pairing and spread the elite-level damage out more. Given that Mueller probably did better than expected against strong competition, it is not the craziest idea in the world.
As a whole, Butcher should be given much more of a prominent role and much earlier than after when Lovejoy was traded. He has done something that the other Devils defensemen have not done last season: significantly impact the CF% in a good way against all competition levels. Give the young-ish man more minutes, Hynes. He’s earned it.
Lastly, a look at this makes more of a recommendation of what to do with the blueline’s depth. Yakovlev did not stick around in New Jersey and this supports why he did not. A third-pairing defenseman should do well against his peers, not get buried in attempts. Santini has been a killer of the Devils’ offense. So while Carrick has not been much of a needle-mover himself, he has been better than Santini. Unless Santini shows he can help support the offense and do more than just skate around in his own end of the rink, then Carrick should have an edge on making the lineup over both. Of course, given how Carrick’s season went, I could understand it if the Devils sought out another defenseman entirely for that third-pairing RD and/or seventh defenseman role.
Now I turn to you. What do you make of the WoodMoney method approach to quality of competition? Did this post make you appreciate what Butcher and Vatanen did last season? Would you agree that Severson has been better than Greene, even if both were not at all that good from a Corsi perspective? Do you now agree (or agree further) that Carrick is a better defenseman today than Santini? Please leave your answers and other thoughts about the findings in this post and the Devils defensemen against levels of competition in the comments. Thank you for reading.