State of the New Jersey Devils Defensemen in Early November

Good times are at hand for the New Jersey Devils. The season has surpassed the beginning part of the season (we’re over 15% in) and the Devils are in a good spot. They’re 8-3-3, they’re sitting in a playoff spot as of Sunday morning, Taylor Hall has been the offensive stud as hoped for when he was acquired, and the team has provided a lot of great highlights in their first fourteen games of the season. They are about to embark on a difficult road trip but they could not ask for a better run of results heading into it. What’s more is that they actually do look fast, attacking, and supportive on the ice. If you go to Corsica, you’ll learn the team has a shot for per sixty-minute rate (SF/60) of 31.82 - a big improvement over last season’s paltry (and league-worst) 23.86 SF/60. There’s plenty to like about the 2016-17 New Jersey Devils. Except for perhaps the defense.

While the Devils have been more prolific at shooting the puck, they’ve allowed more shots than before in recent memory. Corsica goes as far back as the 2007-08 season. The Devils have not had a shots against per sixty-minute rate (SA/60) higher than 30 in any of their last nine seasons. The 2016-17 Devils currently have a SA/60 of 30.88. Attempts against are also up, from 50.20 attempts (Corsi) per sixty minutes (CA/60) last season to 54.07 CA/60 - which is also the highest ever recorded for New Jersey at Corsica. What’s more is that while the Devils have a positive shots for percentage rate, they’re still out-attempted as a team in 5-on-5 play with a Corsi For% of 48.15%. While that’s better than the two previous seasons, it’s still in the red and it also points to the Devils having issues on defense. Defense was a concern heading into this season given that Adam Larsson was moved for Taylor Hall, David Schlemko signed with San Jose, and it was not clear how much Ben Lovejoy, Kyle Quincey, and Yohann Auvitu would contribute as they were all new to the team. Combined with a forward group that may not really be considered to be strong on defense, it appears there are issues given the rise in attempts and shots against in 5-on-5 play. That’s all the more reason to look at the on-ice 5-on-5 stats of the Devils’ six defensemen to determine the state of this defense.

Fortunately for the purposes of this comparison, the New Jersey Devils have only used six defensemen this season. Jon Merrill has been out with a broken finger since the middle of preseason. Steve Santini and Vojtech Mozik have been on the roster, but did not appear in a game. As a result, all six defensemen have played in every game and appeared in well over a hundred minutes in 5-on-5 play. In order to provide perspective, I pulled the data from all defensemen who have played over a hundred minutes in 5-on-5 situations and determined the median out of 179 players. So the following charts will not just highlight how the Devils defensemen stack up among themselves; there will be a “Median NHL D” to compare them to. All data in this post comes from Corsica unless otherwise specified; headings that are in italics was how the table was sorted.

Corsi & Shots

All data from Corsi.ca

In terms of shots against per sixty minutes by player, the median NHL defenseman posted up a better rate than every Devil except for what was used as the initial third pairing: Yohann Auvitu and Kyle Quincey. A closer look at the numbers show that while Quincey may be present when the Devils are out-shooting their opponents, the Devils are also out-attempting their opponents and out-shooting their opponents more when Auvitu was present. It speaks to how much more offense Auvitu can bring to the table.

In any case, the “top four” are all on the wrong side of 30 SA/60. It is weird to see four Devils defensemen break 30 SA/60 now when only John Moore surpassed that mark last season. With that in mind, it is not exactly surprising to me to see Moore rank last in SA/60 and even CA/60 on the team. If you want to point to who’s been the least effective on defense, these numbers suggest it is Moore. While Damon Severson’s SA/60 rate is not too much lower than Ben Lovejoy’s, his CA/60 is closer to Andy Greene’s. That speaks to him being more effective than Lovejoy or Moore. It also helps that the other players with Greene-Severson have been supportive with their attack as they both have a superior CF/60 and SF/60 than Lovejoy and especially Moore. Still, the performances of the Greene-Severson and Moore-Lovejoy pairing are more at the core in the team allowing more shots in 5-on-5 play so far in this season.

Context

Of course, we need some context to go with these figures. CA/60 and SA/60 alone are not enough. Let’s look at their usage first:

All data from Corsi.ca

ZSR is Zone Start Ratio, or the percentage of offensive zone starts to defensive zone starts. Auvitu is the only defenseman who has received more offensive zone starts than defensive zone starts in 5-on-5 play. The coaches recognize he has an offensive skillset and have used him as such. The other five defensemen, well, they’ve started in their own end more than the Median NHL Defenseman. Greene and Lovejoy have had more difficult starts early on than the other three sub-50% ZSR defenders.

From a standpoint of quality of competition by time on ice (TOI QC), it appears that Greene and Severson have received the tougher competition. They are the only pairing with a higher TOI QC than Median NHL Defenseman. They also have a lower Corsi For Quality of Teammates (CF QT) than Moore or Lovejoy. Yet, it is not a clear picture. According to Corsi For Quality of Competition(CF QC), Moore and Lovejoy have had tougher matchups and they have a lower - and in Lovejoy’s case, the lowest - time on ice quality of teammates (TOI QT) than those two. The only real clarity here is that Quincey and Auvitu have largely been used as a third pairing. The top four is muddled between who really has it harder between competition and teammate utilization.

That be a result of the Devils’ schedule. They’ve played an even split between home and road games. That means the Devils had the luxury of using Greene and Severson to go against the other team’s best and play with the Devils’ top forwards at home half of the time. It also means that opponents had more control on who Greene-Severson and Moore-Lovejoy matched up with.

Of course, this is all on-ice context. Let’s look at some more. Let’s look at what they have been present on the ice for in terms of goals.

All data from Corsi.ca

This chart here really highlights why the coaches may have kept with the same six defensemen in the lineup. For as many shots and attempts against Moore has been present for, he’s only been on the ice for only three goals against. For the issues in Quincey’s game, he’s been present for only four goals against. Both have been present for more goals than not. Lovejoy? Tied with Severson and the Median NHL Defensemen. In short, while their shots and attempts against rates are not encouraging, it has not damaged the Devils.

Rather, not yet damaged them. Moore, Quincey, and Auvitu have high PDO values (sum of shooting percentage, Sh%, and save percentage, Sv%). The rule of thumb with PDO is that high values tend to fall over time and low ones (look at Greene) tend to rise. It is not apparent what, if any, connection a defenseman has with shooting or save percentages by a team in 5-on-5 play. It is apparent that Moore has been especially fortunate that the goalies have been utterly fantastic while he is on the ice. I fear that it’ll take more goals against for Moore (or Quincey, or Auvitu) to sit any time soon. Of course, the seventh defenseman option has been a rookie so far this season when one was available. A more experienced Jon Merrill may lower the requirement to be scratched for a night when Merrill is re-activated onto the roster.

Expected Goals and Scoring Chances

Now, there are two other ways to look at the performances of the Devils defensemen primarily through Corsica. The first is Manny Elk’s Expected Goals stat. You can read up more about it at Corsica. The general idea is that what kinds of shots (for or against) can drive to how many goals should be scored and that is what Elk has modeled. How many goals against were expected to go against the Devils defensemen?

All data from Corsi.ca

The answer: not many. The Median NHL Defenseman has a expected goals against (xGA) of 8.93 and an expected goals against per sixty minutes rate (xGA60) of 2.36. Only Lovejoy has had a higher expected goals against among the Devils. Also, only Moore and Lovejoy have higher xGA60 rates. This stat puts something behind the notion that I have about Greene and Severson. While their SA/60 rate above 30 was surprising to me, I regarded their play as generally effective and the xGA60 and xGA rates reflect that. My understanding is that what they’re allowing are not likely to go in all that much relative to the Median NHL Defenseman. At the same time, the more contentious play of Moore and Lovejoy has yielded higher expected goals against rates. Fortunately for the Devils, those expected goal rates have not borne out into actual goals against. Per the previous chart, Moore has been on the ice for a mere three 5-on-5 goals against New Jersey and Lovejoy has only been present for seven. For their sake, may they continue to surpass expectations.

As an aside, this chart also shows that Quincey and Auvitu really have not been hampering the cause on defense. While Quincey may have had bad moments, he has the lowest expected goals against rate on the team. I did not expect that We know that their usage has been easier than the other two pairings. But that is a benefit in that the third pairing has not really killed New Jersey by these stats so far.

The second stat? Scoring chances. The hope is that the defensemen would protect the most dangerous parts of the offensive zone. Defenses are designed to at least try to keep the opposition to the outside. I like the concept of counting scoring chances; I’m slowly more accepting their automatic counts at Corsica, Natural Stat Trick, etc. So let’s look at that and KPO%. KPO% stands for Keeping Pucks Outside percentage. It was developed by Andrew Kerison at Original Six Analytics and I think it’s worth highlighting. Intuitively, it makes sense as it’s the inverse percentage of scoring chances against over shot attempts against. Here’s how the Devils defensemen fared among chances.

All data from Corsi.ca

While the Devils, as a team, may have a higher shots and attempts allowed rate than in previous seasons, they’re still stingy at allowing scoring chances. John Moore has been present for the most scoring chances on the team and that’s still just better than the Median NHL Defensemen used for this analysis. He’s also better in KPO%; keeping so many pucks to the outside. I look at this chart and conclude that if there is one thing that the Devils do well, it is this. Assuming that chances are being counted correctly. It also in this light that Greene, Lovejoy, and - surprising to me - Quincey has been excellent in this regard. If nothing else from this analysis, I have a more positive opinion of Quincey’s overall play.

By the way, if there is a negative here, it’s on the for side of the stat. As in, these defensemen are not present for many scoring chances generated by the Devils. As defensemen, they’re not likely to be the ones to do be the ones taking the scoring chance. That is understandable. That they all have low scoring chance for per sixty minute rates speaks to how the other team has been successful at keeping the Devils to the outside in 5-on-5 play. Here’s hoping that improves as the season goes on.

Production

Speaking of offense, let’s look at production to close this deeper dive into the blueline’s 5-on-5 play in this young season.

All data from Corsi.ca

Severson is the not only the defensemen’s leader in 5-on-5 scoring, but he is second behind Taylor Hall in points with eleven in all situations. That he has been this productive really puts a spotlight on what he has done this season. And that’s good. It mitigates the other aspects of this game, which have not been that bad in other lights. Does the same apply to Moore? Again, being productive will help keep him in the lineup and with four points and eighteen shots in 5-on-5 play, that it does. The other defensemen have chipped in. Auvitu has been the shooter of the group and Lovejoy has been a surprise with eighteen shots on net. Both will get points in time provided they keep that up. This chart shows that Quincey really is not much of an offensive player, although he scored a lovely goal against Buffalo on Saturday.

General Concluding Thoughts

Yes, the Devils have only played fourteen games in this season. However, if there are players struggling on defense, then it does little good to wait to identify them. That said, the Devils have been both more attacking and more allowing of the opposition to attack back in 2016-17 compared to past seasons. Their rate stats for shots and attempts against are higher than ever. A deeper dive into the six defensemen shows that the top four group of Greene, Severson, Lovejoy, and Moore have been at the heart of this increase.

Auvitu and Quincey have been mostly used as a third pairing in these fourteen games. The stats show that Auvitu has been the offensive one and Quincey has been more or less a defensive-minded defender. In the larger picture of these games, they have been more than fine given their usage. The PDO for both of them are high, but their other rates make it look fine. Quincey doing so well in terms of SA/60, CA/60, scoring chances against, and expected goals against was all a pleasant surprise. He has had some rough shifts and periods, but over all of these games, he has not been a liability. That’s good.

What is concerning is the play of Moore and Lovejoy. Severson and Greene are both above a median NHL defenseman with 100 minutes of 5-on-5 play in terms of CA/60 and SA/60. But that is mitigated by their low scoring chance rates and low expected goal values. Further mitigation comes from Severson’s production on offense and Greene starting so many more of his shifts in his own end. Moore and Lovejoy have been leakier than Greene and Severson and that is where the concern lies. It has not been too damaging in terms of goals against and especially in terms of scoring chances against. All of the Devils defensemen have been good at keeping the puck to the outside so far in this season - even Moore. However, the concern is that once the goaltending stops being amazing when Moore is on the ice, Moore’s play will start to be damaging - even in spite of his good production so far in 5-on-5 play.

Can the Devils make any changes? Recent games have seen a switch in defensive pairings. According to the three most recent games listing of defensive pairings at Left Wing Lock, Moore has been mostly paired with Quincey and Lovejoy has been mostly paired with Auvitu. It remains to be seen whether they will stay, but it is a sign that the coaching staff has seen enough of Moore-Lovejoy as a pairing. Jon Merrill is closer to returning to the lineup. While Merrill has not exactly been good, he is experienced and he has had at least some better underlying numbers than Moore last season and several of the current Devils. I’m not saying Merrill would be immediately great, but he may stem some of the issues. I’d like to think he could replace the team’s most problematic-on-paper defender, Moore, for a few games to see how he’ll fit in. I think it is more likely for Quincey or Auvitu to take a seat to start. We’ll see how the coaches approach it soon. After all, Merrill did practice at the end of last week; it won’t be long before he is activated.

Overall, it appears that the Devils “top four” could be better as they are the ones soaking up much of the increase in shots and attempts against the team. The kinds of shots and attempts mitigates how bad that increase has been. The hope is that the Devils’ blueline can remain relatively stingy at preventing scoring chances and work to cut down on attempts and shots in general. It would not surprise me if the numbers are similar by the time 2017 comes around (or 23 games from now) and the general thought comes about that the Devils could use help on defense.

One last thought: Moore has had the highest SA/60 on the team last season and he’s the highest right now in this season. I think it may be worth taking a closer look at who he has been playing with to see what other factors are contributing to that figure. That is something for another post on another day.

Thank you all for reading through this deeper dive into the Devils’ defensemen. I know it is still early in this season, but I believe that even just fourteen games of data highlights how the defense has been performing and justifies some of the concerns fans may have it. We’ll see whether anything is done about them as the season goes on; hopefully, it is all for the better in time for the good of the team. Please let me know what you think of the Devils defensemen in 2016-17 in the comments.