After the New Jersey Devils lost to the Islanders, 3-4 in overtime, I had plenty of thoughts and one question. I wrote a recap to get the thoughts out there, as usual. The one question deserved its own post. Are any of the defensemen on the 2018-19 New Jersey Devils good?
I am confident I am not the only Devils fan to ask this question. I am confident there are fans who had that question before this season even began. This is an organization who had a dominant blueline at its peaks when they were trapping chumps, winning Cups, and making many in the hockey world moan and whine about the Devils. However, those years are well in the past. It was a defensive corps with two Hall of Famers (Stevens, Niedermayer), a lock for the Hall of Very Good (Rafalski), and other solid defenders (White, Driver, Daneyko before he got old, etc.). That level of talent isn’t on the 2018-19 roster. But the game in the National Hockey League has also changed from the late 1990s and early 2000s too. We must ask this question with respect to current times and current defensemen. And so we must at least compare the current group of Devils defensemen with their current peers.
Micro vs. Macro
Before jumping further, there is a larger philosophical question to address. What makes a player good? In the heat of a moment or in the short term of a few games, we would expect a good player to succeed at what they do and not fail miserably in the process. For a defenseman, those failures would include coughing up the puck in their own end or not stopping a player from collecting their one rebound in a one-on-one situation. Those are micro situations. Events, if you will. We can agree that a player can have a bad event. And those with particularly good memories will remember specific events over anything else, even good events like scoring a goal, helping the goalie prevent one, creating a breakout, and so forth.
The macro situation looks at multiple games, such as the season so far. A player with several bad events or a lack of good events will show up in a macro view. Such as that when they are on the ice, the team sees an increase in shots or scoring chance against. Or that they are not contributing to offensive plays so the play often is in their end of the rink.
My take is that any defenseman of any quality at any level will have good and bad events. Like any other fan, I am bothered by the bad events and pleased with the good ones. However, if a player is truly good (or bad), then they I want to see a macro view rather than focus on a micro view. So even though one game and a couple of bad plays is driving this post, I need a larger pool of information and see how the Devils defensemen have done this season to really say if they are good or not.
This is a big part of why I use stats the way I do. So, yes, Defenseman X made a bad play one time and I wish he didn’t make it either and I’m glad you remember it, but a one-time error does not undercut an entire season. If he is as bad as you may think, then we’ll see it in the stats.
I will also need a point of comparison from the whole league. Prior to Saturday’s games, I pulled data from Corsica.hockey and Natural Stat Trick for the Devils defenseman and identified the league median. Conveniently, 236 defensemen played at least 50 minutes of 5-on-5 hockey this season. Therefore, the stats for 118th place at each site has been combined for a made-up player named League Median. I’m also focusing on 5-on-5 play because the Devils do not use all of their defensemen for the power play or the penalty kill. I want an even playing field, so to speak, and so I want to see how the defenders have performed in the most common situation in hockey. A good player would perform well in the most common game situation and a bad player would perform poorly in the most common game situation. Besides, I do not think anyone is concerned about the Devils’ penalty kill and I do not think the power play issues are defenseman-specific.
Now that you know where I’m coming from and how I am doing it, let’s get into the data. Seven Devils defensemen made the 50-minute cut-off: Andy Greene, Damon Severson, Sami Vatanen, Mirco Mueller, Will Butcher, Ben Lovejoy, and with 65 minutes, Egor Yakovlev.
I’d take Yakovlev’s numbers with a grain of salt given his very low amount of ice time; thus, I will focus on the other six defensemen who are the team’s regulars. I will offer this one general comment about Yakovlev. There’s enough good things in those 65 minutes and change that warrants some further action from Yakovlev - which he received in Tampa Bay last night.
The Devils Defensemen vs. League Median
Numbers in green are above the league median, numbers in red are below it, and numbers in black are just pretty close to it. Cells are highlighted to draw your attention to some impressive or terrible values.
There is a lot of stats here so let’s go through them bit by bit.
Game Score - Game Score was developed by Dom “omgitsdomi” Luszczyszyn, which is explained in this post at Hockey-Graphs back in 2016. Basically, it’s a single number that takes into account a player’s whole performance in a game. Positive is better and Corsica tabulates it over a season.
On the New Jersey Devils, Damon Severson is your runaway leader in Game Score. Whether it’s in the raw category or in the per-60 minute rate, Severson has contributing the most in 5-on-5. This should not be a huge surprise since he is also the runaway leader among the Devils defense in 5-on-5 scoring with two goals and eleven points. Vatanen is second with five points and he comes second in this category. He’s at least above the league median.
The only defensemen who come up short here are Andy Greene and Will Butcher. They are not too far away from the league median. However, Game Score suggests that Greene’s rate of contributions could be much better.
Individual Shots, Goals, and Point Rates - Defensemen are a part of the offense and production is a great way to contribute. The Devils defense as a whole has been lacking in 5-on-5 goals. Severson and Butcher entered Sunday’s game as the only ones with goals and Severson is leader with two.
This has not deterred other defensemen from firing away. While Greene and Mueller are shooting below the rate of a league median defensemen, their partners Severson and Sami Vatanen have exceeded them. That may explain why the rates are different. However, Vatanen’s shots per 60 minute rate is far and away higher than rest of the defense. Surely, the goals will come for him. Curiously, Ben Lovejoy has a higher shooting rate than Will Butcher. While I like the initiative, I’d rather have Butcher take more shots. Lovejoy has his uses but shooting is not really one of them.
In terms of points, Butcher and Greene are lagging behind the rest. In contrast, Severson and Vatanen have excellent rates of points. Vatanen’s could be higher just from all of the rubber he’s firing. Severson just needs to keep up what he’s doing and the coaches just need to keep playing him with scoring forwards.
Corsi (Attempts) - In terms of Corsi For% (rate of attempts by the Devils over the rate of total attempts by both teams), the breakeven mark is 50%. While the defensemen are not the sole reasons why they have a high or low CF%, their play will contribute to it. The good news is that majority of the Devils defense is above 50%. The bad news is that Andy Greene and Ben Lovejoy are below 50%. This is a bit odd because their regular partners are above 50%. Nobody is really high in this category but for the most part, in this macro view, the Devils defensemen are not being wrecked. Greene is the low man at just below 49%, which is not too terrible and a far cry of where he was this time last season.
In terms of the rates that contribute to these percentages, the team has an amazing Corsi For per 60 minute rate with Vatanen on the ice. A rate over 61 is excellent. His partner, Mueller, and Severson also rate highly. The entire blueline is above the league median, in fact. Only Butcher is “low” but he’s still ahead of the median mark. This suggests that the defense has been helping for the most part in terms of creating attempts.
The issue is with the opposition. The Corsi Against per-60 minute rates are not as good across the board. Greene’s CA/60 is above 60, which means that he’s getting to play lots of defense. Vatanen’s CA/60 is also rather high; he comes out ahead in CF% because the Devils many more attempts when he’s on the ice. It could be that Vatanen is a relatively high-event player. Lovejoy and Severson are also above the league median in this category, which is not good but not to the levels of Greene or Vatanen. The only regulars who have a CA/60 below the league median are Mueller and Butcher. Butcher’s an interesting standout since he’s not regarded for defense. Perhaps he is doing better than he is given credit for?
I added relative stats, which are the difference in the team’s stats when that player is on the ice from when that player is off the ice. Ideally, you’d like to see the CF% improve so positive numbers are good and negative numbers are bad. By relative CF%, Mueller and Vatanen have been the most positive. In other words, when 25-45 step on the ice, the Devils have improved in taking more attempts than their opponents. That is a good thing. Severson is just above positive, Butcher is just below negative, but their partners are both deeper in the red. The team’s CF% has dropped when Greene or Lovejoy step on the ice in 5-on-5. How it has not hurt their partners so much, I am not sure.
Shots - The shots for percentage (SF%) for each defensemen sort-of follows the CF%: Greene and Lovejoy are below 50% and everyone else is positive. The good news is that Greene and Lovejoy are a bit better in this regard; they’re within a percentage point of breaking even. Plus, Mueller and Vatanen are both higher. Mueller is especially good; when he’s been on the ice this season, the Devils have taken just over 56% of the shots happening. That’s really good.
Mueller’s rate stats are also really good. The team’s shots for per-60 minute rate is above 36 when he’s on the ice. That Mueller-Vatanen pairing has been doing something right when it comes to shots as only Vatanen has a higher SF/60. Maybe it is more appropriate to say that the coaches have been doing something right given who could be in front of that pairing. The point remains: you want shots, you want 25-45 on the ice. Amazingly, all six regular defensemen beat the league median in SF/60. The lowest one is Butcher and he still is ahead of L. Median by at least one SF/60.
The bad news is, again, in shots against per 60 minutes. Almost all of the regulars are above of the league median. The worst of the bunch is Greene, although his partner, Severson, is not too close behind. Lovejoy and Vatanen are above 31 SA/60, too. The exception is Mueller. I did a double-take, but I have confirmed it. His SA/60 rate is really that low. He is the only defenseman with a SA/60 below 30 on the team right now. Is Mueller deserving of more credit? This suggests so.
In terms of relative SF%, the Mueller and Vatanen pairing stick out in a great away. When they have stepped on the ice, the Devils have had a better share of shots. When the other pairings step on the ice, their share is reduced. It is especially reduced when Greene and Lovejoy step on the ice. While Severson and Butcher have negative SF%s, their partners have been worse. I’m wondering if their defensive styles lend themselves to being on defense more often - which may hold the team back a bit.
Goals & Expected Goals - I do not think that Devils defensemen should necessarily be credited for a high goal scoring rate or punished for a high goal against rate. However, it is a factor in determining whether things have gone well or not. To add another layer, I’ve included Expected Goals. The Expected Goals model looks at all shot locations when a player is on the ice, weights each one, and determines how many goals theoretically should have been scored. It is a good way to at least see if someone has a rate well above or below expectations.
For the Devils, the leading pairing in actual Goals For% is Lovejoy and Butcher. Teams have simply not scored a lot on them and they have been present for plenty. Butcher has the highest team on-ice shooting percentage among defensemen and Lovejoy is not too shabby at 9.09%. Both defensemen have had excellent goaltending behind them; particularly Lovejoy, who has benefited from a 95.16% save percentage.
However, the expected goals model suggests that they have been fortunate in this regard. Based on where and how many shots have been allowed, the model has a higher xGA/60 than their actual GA/60. In other words, do not be so stunned if they start being scored upon more in the future. This is particularly bad for Lovejoy as his expected GF% is the only one on the blueline that is below 50% and his actual GF% is an incredible 64.7%.
The model suggests that better times may be ahead for the other two pairings. While Severson’s actual GA/60 is pretty high, he has the highest GF/60 and the only such rate on the team above 3 goals. Greene’s rates are not as extreme, but he comes out ahead as well. The expected goals model suggests the proportion can be even better. The pairing that has been run over by goals in reality is Mueller and Vatanen. Despite being present for so many shots and attempts taken, they have not turned into goals. Their on-ice team shooting percentages are less than 6%. That can (hopefully) turn around in the future. But this means that their rates of goals against yield a GF%s in the 43% range - which is not good. That’s at a point where a coach may want to switch things around instead of seeing a pairing get lit up. Again, the expected goals model suggests better things for them. While Vatanen’s xGA/60 being above three means that there is too much being allowed; his xGF/60 is also above three and is higher so at least the model thinks he should come out ahead. That the values are so high for Vatanen and not Mueller may suggest a switch in that regard.
The relative stats for goals for percentage are what you would expect given the actual percentages. Lovejoy and Butcher have equalled big gains; and Mueller and Vatanen have equaled big losses. At least the Greene-Severson pairing is positive in this regard, even if it is by a smaller magnitude. The expected goals model is not as kind from a relative perspective. Greene and Severson are the only ones to provide positive impact per the model. The other two yield drops in this category, even if the model thinks Mueller and Vatanen should not be as scored upon as they have been, they have not stingy enough.
Penalties - A quick note on penalties taken and drawn. I do not expect defensemen to draw calls since they’re usually the ones initiating contact. Plus, I do think they do foul players to prevent potential scoring plays. They also at risk of fouling players when they’re beaten too. Still, I included them to show whether anyone has been particularly good or bad at it. While Severson and Vatanen have taken five penalties each - which is above the league median - they also have drawn almost as many in response. That is quite good given their positions. The one who sticks out like a sore thumb here is Lovejoy. He leads the defense with seven penalties in 5-on-5 and he has drawn one. Watch your stick and what you do out there, Lovejoy. The calls are not helping. Since he’s usually on the first penalty killing unit, those calls also deprive him of a chance to contribute.
Now, while the Expected Goals model does take location into account, that is not necessarily the same as identifying how the players have done with respect to scoring chances or high-danger (slot and crease) scoring chances. Natural Stat Trick has this data explicitly, so here is the data compared with the league median.
From a scoring chance (SC) perspective, half of the defense looks good while the other half does not. Butcher and Severson are the leaders in SCF%. They both have scoring chance against per 60 minutes rates well below the league median and they have scoring chance rates per 60 minutes by the Devils well above the league median. This is not to say they do not give up scoring chances or only create them, but they’re way better at it than the others. Plus, when they step on the ice, the team’s proportion of scoring chances goes much more in the right direction.
Their partners have not been as good, but they are not total disasters. Greene has been good from a SCF/60 standpoint but he’s high in the SCA/60 rate. Lovejoy is just above the league median in SCA/60 but his SCF/60 rate is pretty low. Greene is better in terms of percentage and relative SCF%. Lovejoy, not so much. At least it has not deterred his partner.
The pairing that really struggles is Mueller and Vatanen. Despite all of those shots and attempts that happen on the ice, they’re not necessarily chances. Mueller’s rate of scoring chances for is the lowest on the team. He is at least not giving up a ton of chances more than the league median. Vatanen has a higher rate of scoring chances for - and high scoring chance against rate. The highest among Devils defensemen. This is further evidence of how high-event Vatanen has been and that it has not always been a good thing from a defensive standpoint.
I’m convinced that the Devils, as a team, focuses on high-danger scoring chances (HDCF). While scoring chances has been a mixed back for the defense, everyone is above 50% HDCF%. Everyone’s on-ice high danger chance rate is above the league median. Even the Lovejoy, Mueller, and Vatanen are above 50%. Severson’s and Butcher’s HDCF/60 and HDCA/60 rates are amazing. Not just as stand alone rates but in how far apart they are from each other. The only negatives here are Vatanen having a high HDCA/60, Lovejoy’s relative HDCF% percentage being so negative, and the not-as-negative relative rates for Mueller and Vatanen.
Goals Above Replacement - Goals Above Replacement (GAR). This is another model attempting to provide comprehensive value for a player. This is an orange in a bowl full of apples as this is not just for 5-on-5 play but all situations, including special teams. It also compares against a replacement-level skater, not the league median. Still, it is good to have some variety in the fruit. CJ posted a graph of everyone’s GAR, as determined by @EvolvingWild (Josh and Luke), in his post on Wednesday morning. The chart is available at Evolving-Hockey; here’s the chart prior to Sunday’s game.
By the twins’ GAR model, only Vatanen has provided negative value. Mueller’s even strength value is also negative, but he is salvaged by his PK play and penalties (Mueller has drawn three, taken two). The other defensemen are better than a replacement level player, with Severson being the third most valuable skater by this model. I would not treat GAR as the be-all, end-all stat; but it provides further evidence of who is doing well and who is not doing well from a macro perspective.
Conclusions: Yes, There Are Good Defensemen on the 2018-19 Devils
To answer the initial question, there are defensemen on 2018-19 Devils who have been good this season. No one is perfect and no one will be confused for being an all-star level defenseman. But there are good players on this blueline.
The best among them would arguably be Severson. Yes, Severson. He’s been productive when the other defensemen have not been. He’s drawn as many calls as he’s taken. When he’s on the ice, the Devils are still out-attempting, out-shooting, and out-scoring their competition. Severson has been great from a scoring chance and a high-danger scoring chance perspective. He is not over-achieving expectations. With the exception of shots for percentage, when he steps on the ice, the Devils’ percentages improve whether it is for attempts or scoring chances.
I can already foresee someone complaining about how this could be when he failed to cover an open skater in a 2-on-1 in overtime or how he had a bad turnover in Carolina in Sunday. I refer you back to the Micro vs. Macro section. Yes, these were bad mistakes. However, he is not making bad mistakes constantly enough or consistently enough to really hurt the team. Whatever the team has allowed with him on the ice, they also have generated even more when he’s on the ice. Severson can have issues in situations, but he is not the trainwreck you may perceive him to be. And, honestly, I do not think anyone is a true trainwreck here.
Also falling into the good column would be Will Butcher. He has not impressed in some areas. It is somewhat of a concern that he has not been productive and that Ben Lovejoy has taken more shots than him. Still, Butcher has not been worse than the league median except for a handful of categories. While not producing much is one thing, he has not been a sieve, a pylon, or a body on the the ice this season. We can expect more, but he has not been bad at all either.
Your mileage may vary on the other defensemen. The worst of the bunch is Vatanen and that may be more of a function of how high-event he has been than actually being bad. When he’s on the ice, the Devils are shooting and attempting shots at a high rate. The issue is that the other team has been shooting, attempting, earning scoring chances, and scoring goals at high rates. He and his teammates have not generated enough goals to offset the negatives too. Even with that, he still fires a lot of shots, he has multiple points in 5-on-5 already, he’s almost drawn as much as he’s taken in penalties, and if you want a highly offensive team, then that has happened when he’s on the ice this season. His partner Mueller has been far better with CA/60 and especially SA/60 rates, so I wonder if the pairing is an issue more than the individual defensemen. As for Greene and Lovejoy, again, while they have some sore spots, they have their positives too. Arguably enough to justify their use.
From the macro perspective, no one is truly terrible or across-the-board worse than the league median defenseman. There are some poor spots here and there, but I did not have to make an entire column red for anyone.
The good news is that the coaching staff appears to be more comfortable with Yakovlev to start putting him in more games. For example, he replaced Mueller for yesterday’s game. To help account for those events - the micro view - where a defenseman played poorly, having a spare ready to step in is a plus. Yakovlev can be that one for the left sided defensemen. When Steve Santini is ready to play, he could very well be that one for the right sided defensemen. I do not know if head coach John Hynes and general manager Ray Shero want to carry eight defensemen on their active roster. But they can and it can stem any “bleeding” that comes from a poor run of games.
I am also coming around to the idea that perhaps the pairings should be switched up. From this exercise, I’m getting the sense that Mueller and Vatanen may not be working so well together and that both Severson and Butcher could be better with different partners. I do not want to relive the Greene-Lovejoy experience. But I wonder if a return of Greene-Vatanen and Mueller-Severson would be best for all involved? Or possibly moving Butcher up into a more prominent pairing with Severson while Mueller and Lovejoy can go be a more defensive third pairing? The coaches have options. While, again, none of the 2018-19 Devils defensemen are bad; there is room for improvement - in both the micro and macro perspectives.
I now put the question to you. Are any of the defensemen on the 2018-19 New Jersey Devils good? If so, why do you think that way? If not, why not? Please leave your answers and other thoughts about the Devils defensemen in the comments. Keep in mind this was all written before the Tampa Bay game, so if anything good/bad happens there, then it is not in the post. Thank you for reading.