Defense is an important aspect of the game of hockey. As much emphasis is placed on pushing the play and generating offense, good teams need defensemen to limit, corral, and deny the opposition of offense of their own. Success for them is making sure the other team has as few of those as possible. As a result, it's not always easy to identify who's done well on defense. They don't make highlights unless A) they do something on offense, B) they make a spectacular play, or, worst of all, C) they make a mistake that directly or indirectly leads to a goal against. Holding position against an on-coming puck carrier, clearing out a dump-in attempt, or knocking away a loose puck after a save are all important and may draw cheers at the moment it happens. However, there's no simple, countable stat (yet?) that catches those sorts of plays. Trying to recall it later is the problem and so mistakes take precedence.
The introductory preface is there because the New Jersey Devils defense has been one of the best in the league in terms of shot prevention and I can understand if your initial reaction is somewhat skeptical. Even I have a bit of an issue typing it. The Devils looked all kinds of sloppy for periods, if not games, in their own end early this season. There have been goals against directly attributable to the defense. The individual names on the blueline don't jump out as top defenders.
However, the numbers don't lie and the Devils really have been one of the stingiest teams in the NHL when it comes to shots against, one the measures of whether a defense has been successful as a group. In general, they are third in the league in shots against per game with 25.8 prior to Wednesday night's game. They've only allowed 30 or more shots in three of their thirteen games. Five-on-five is the most common situation for teams in games and the Devils are currently third in the league in shots against per sixty minutes, 25.0, according to Behind the Net. That all points to a team defense that usually keeps the opposition at bay.
This raises the question, if the team has been solid in their own end, then who's making it happen? Who's getting the tough assignments and how are they doing at five-on-five, the most common game situation? While the team has played only thirteen games, the advanced five-on-five stats at Behind the Net make it pretty clear how Peter DeBoer has set the defense:
I've organized the numbers by Corsi Rel QoC, or the average relative Corsi* of the opposition the player faces. Essentially, it's a measure of how strong their opposition plays against; higher is tougher. How the Corsi Rel QoC breaks down coincides with how each defenseman has been used. You will notice that Peter Harrold and his one game isn't included. *(Reminder: Corsi is the summation of all shooting attempts - shots, blocks, misses, and goals - for and against.)
For the Devils, the defensemen with the toughest assignments so far has been Bryce Salvador. The captain has played the most at evens (see time on ice per 60, TOI/60) and he usually starts in his own end of the rink (41.5 offensive zone start percentage or OZS%), so he's seen a lot of opposition's better players. He hasn't done so well at it. His on-ice Corsi is the lowest on the team as his on-ice shots against per 60 (SA/60) rate. This means when he does take shifts, he's usually forced to make plays defense at even strength because the opposition is attacking. When he comes off the ice, the opposition isn't as successful, as evidenced by his low off-ice so The good news for the captain is that it's only led to six goals against, thanks to some strong goaltending behind him.
However, there may be better days ahead for #24. In recent games, Peter DeBoer has paired him with Mark Fayne. Like Salvador, Fayne plays against a strong level of competition. Unlike Salvador, the puck usually goes in the right direction when he's on the ice. He's been playing with Salvador for the last week or so and his on-ice Corsi rate remains the highest among the top four. He's benefited from even better goaltending when he's on the ice; but his on-ice and off-ice SA/60 rates suggest that Fayne has a very positive effect on the defense. When he has been on the ice, the opposition is severely quelled. Not only that, but Salvador's own on-ice Corsi rate has improved a bit (for example, it was below -10 before the Carolina game). We shall see how long this duo can keep it up.
Speaking of duos, Adam Larsson and Andy Greene have been put together before the winning streak and it looks like a success so far. They represent a second pairing given that their Corsi Rel QoC is lower than Salvador and Fayne while higher than zero. DeBoer has trusted them against good players, as both players have seen increase in this category in recent games. They also have started far more in their own end than in the opposition's end - especially for Larsson - which is further proof that the coaches trust this pairing. Supporting that decision further is that Greene and Larsson have on-ice SA/60 rates well below the team's rate of 25 in 5-on-5 play. However, there are a few downsides. The play hasn't gone forward often when they're on the ice at evens; it has gone against New Jersey more often. And the Corsi rate jumps up when they get off. Since their SA/60 rates are so low, I wonder if it's mostly blocks or misses. The second downside is that they've been really helped out by some high on-ice shooting percentage. That's not going to last. It may not be a big issue from a defending standpoint, yet it suggests the pairing doesn't bring much offense to the table.
Lastly, there's the third pairing of Marek Zidlicky and Anton Volchenkov. Zidlicky didn't start on the third pairing but he's found his way down there with his performances so far. Volchenkov has been a mainstay from last season in that spot. Both are essentially specialists; they get far fewer minutes at evens but their overall time on ice has been boosted through special teams. Zidlicky is the first choice on the power play whereas Volchenkov is on the top pairing on the penalty kill. Keeping both so low has worked so far. Zidlicky and Volchenkov have been able to get good possession based on their on-ice Corsi rates against weak competition (their Corsi Rel QoC values are negative) and favorable zone starts (their OZS% is over 50). Their lower PDO tells me that they aren't due for some regression in luck soon. Their on-ice SA/60 rates aren't bad at all, though those rates are lower when they come off the ice. As a third pairing, they've been fine.
This leaves the seventh man, the odd man out, Henrik Tallinder. It's disappointing to see him regularly scratched now. When he was playing this season, he was on the third pairing which is off-putting since he was a good top-four defenseman in past seasons with New Jersey. Tallinder helped Fayne along as a rookie and done so for Larsson last season when he was healthy. Tallinder has done incredibly well in his limited minutes (that on-ice Corsi rate is ridiculous) and I do believe he'd be better than Zidlicky or Volchenkov right now. However, the team has been very good as of late and the defense has been great at keeping shots against low. Like Larsson a few weeks ago, this means he's not going to get back into the lineup unless someone gets hurt, gets bad, or gets moved.
Now, going back to the group as a whole, five out of the seven defensemen have started most of their shifts in their own end than in the opposition's. Some more than others. Corsi can be easily affected by this, since it's easier to get attempts when starting from an offensive position and harder when starting in the other end of the rink. Salvador, Greene, and Larsson all have negative Corsi values and each have offensive zone start percentages below 42%. We can adjust their Corsi rate for zone starts - and I did the rest for the other defenders. The source for this data is, of course, Behind the Net.
While it's still early enough in the season that the difference between defensive zone starts and offensive zone starts is low, the entire top four comes out looking better. The adjusted Corsi rate went slightly in Fayne's favor and it got Greene above zero and Larsson just below it. Salvador's rate was more than halved. The zone start adjustment dragged Zidlicky real down; perhaps it's further reason why he's on the third pairing. Overall, it means that if the Devils can get more offensive zone starts in general, the defense would look better in possession.
Bringing it all together, there are aspects about the defense that could be better. They remain offensively unproductive. Tallinder has the only even strength goal among all the defensemen this season and everyone else just has one or two assists. Speed and/or an aggressive forecheck can cause them problems, especially against Salvador (who's slow), Larsson (who is still prone to silly errors at times), and Volchenkov (who's also slow). It's still only 13 games so a lot can change in the next two and a half months. And, yes, a sloppy and unfocused performance can do them in as if they were the leakiest defense in the world. However, mistakes happen to every defense. The fact that the Devils regularly keep the opposition to so few shots at even strength means that the errors happen so often that it hurts them.
Simply put, the pairings of Salvador & Fayne, Greene & Larsson, and Zidlicky & Volchenkov have combined for a very solid blueline. While they don't always get the credit or the highlights, they've gotten the job done quite well based on the available metrics. Let's hope they can keep up the good work in general.