![]()
This is the sixth installment of the ILWT Season Preview. In it, I take a look at how we can use deployment and passing data to evaluate the roles the Devils players should be used in. Read on for the details.
When John asked me to write up a preview article on the New Jersey Devils’ coaches and tactics, I started thinking about the data I’d tracked and what it could tell us about how players are deployed, in what situations they are successful, and in what situations they are limited. I put on my hypothetical coaching hat and imagined myself as an NHL coach that looked at this data in between games to get a better grasp on how I could manage in-game situations. If an organization has this data, or their own breed of something like it, I think this is how they should use it.
Starting with the Coaches
Mike wrote up a great piece on the Devils power play, which falls under Assistant Coach Dave Barr’s responsibility . The power play trended up last season and for the reasons Mike stated, it could be a weapon for the team, or it could take a step back. Barr has a few more toys to use in Marty Havlat and Mike Cammalleri, so I'm a glass half-full guy on this.
The penalty kill falls under Assistant Coach Mike Foligno’s responsibility and, again, was a strength of the team last season. John's piece on the PK was thorough and while we can hope to witness another season of solid penalty killing, there's lots of evidence pointing towards some regression. It'll be up Foligno to prevent that as best he can.
Tommy Albelin is the newcomer to the staff, replacing Scott Stevens . I actually really like this move as he spent the last four season coaching the defense in Albany and now he’s going to be coaching many of the team’s young defensemen again. Albelin has spent more time with Eric Gelinas, Jon Merrill, et al than any coach in New Jersey. He’s even spent time with Adam Larsson during his up-and-down Devils career.
Head Coach Peter DeBoer. A lightning rod for most fans, DeBoer does things really well (the Devils system, shot suppression, and goals against) and some things not so well (lineup decisions and situational deployment). First some positives: Justin Bourne did a great job of trying to answer why the Devils are both difficult to play against and also why the number of attempts are so low. While his article addresses the defensive aspects of that question and the style the Devils play to limit chances, the other side of the equation remains a mystery: Why are the Devils unable to attempt enough shots to put them on par with other teams?
Perhaps it’ll be something to look at throughout the season, but for the focus of this piece, I wanted to keep it to deployment, as how coaches use certain players is very much a tactical decision. We’ll look at this in a few phases. First, the zone exit.
The Zone Exit Phase
To start, we’ll look at which players DeBoer used most often in the defensive zone. For this, we can look at their zone starts. Luckily for us, I pulled a lot of data from ExtraSkater before it went dark, so the zone start data originated there.
![Def_zs_pe]()
As always, we start with the defense. We see that Deboer’s first choice for defensive zone face-offs was the pairing of Andy Greene and Mark Fayne. The duo saw a zone start percentage of 47.2% and 47.3%, respectively. Coming out of the zone, however, the duo was not stellar. Only 43.4% of the time did either Greene or Fayne exit the zone with possession (PE%). They weren’t that far behind position leaders Marek Zidlicky at 45.7% and Jon Merrill at 45.3%. Since the coaches openly spoke of their desire to have Merrill fill Fayne’s spot alongside Greene, Merrill may actually improve the zone exit phase of the Devils top defensive pairing.
The problem with DeBoer using "shutdown" defenders like Bryce Salvador (51.3% Zone Starts) and Anton Volchenkov (49.7% Zone Starts) in the defensive zone is that all they effectively shut down is the team’s ability to exit with possession. After Greene and Fayne, Volchenkov saw the next highest defensive zone starts, yet exited with possession only 35.5% of the time—the lowest on the team. Salvador didn’t fare much better as 36.7% of his exits maintained possession. If the goal is to exit the zone with possession, Salvador and Volchenkov are the last defensemen you wanted doing that, yet DeBoer used Volchenkov in the defensive zone more than any other defensemen not named Greene or Fayne, and Salvador was fifth. They shouldn’t have gone near the defensive zone.
Sandwiched between the turnstiles was Peter Harrold (50.8% Zone Starts), DeBoer’s fourth choice for defensive zone starts. Harrold was also the third worst defensemen at exiting the zone (41.2% exits maintained possession). It’s clear that DeBoer and the Devils do not use zone exits and zone starts in this way, as no logical person would hand more defensive zone starts to three players who were the worst at exiting the zone with possession. It’s analogous to working in an office and the person who barely understands computers being tasked with putting together a PowerPoint presentation for new clients. Spoiler alert: it doesn’t end well.
Coaching is partly about managing the strengths and weaknesses of your players. In this situation, DeBoer is heavily using three defensemen in a role better suited to all other defensemen on the roster. The goal in the defensive zone is not to block shots or prevent shots; the goal is to get the puck back and get out of your zone with it.
![Fwd_zs_pe]()
DeBoer, unfortunately, followed up some of his deployment mistakes involving defensemen, by compounding them with some horrendous decisions with his forwards. Stephen Gionta saw the most difficult zone starts (41.1%) and exited with possession 40.7% of the time, worse than any other player not named Mattias Tedenby. If you have a forward that struggles mightily at keeping possession while exiting the zone, why would you use that forward in a defensive role in the first place?
Had the Devils brought Andrei Loktionov back in the fold this season, a line of he, Ryane Clowe, and Tuomo Ruutu may have been a significantly better option at exiting the zone in comparison to the line of Gionta, Ryan Carter, and Steve Bernier. If the Devils have a defensive zone face-off, would you rather have three players that excel at that phase? Or three players that fall on their face?
And yes, teams may wish to deploy their more talented offensive players in the offensive zone more often than the defensive zone, leaving fourth liners like Gionta, Carter, and Bernier to face those more difficult zone starts, but there needs to be a point where coaches ask, "Is this the best approach for my team?"
Mike Sislo, Jacob Josefson, and Loktionov all saw time on the fourth line last season. Each was significantly better at exiting the zone than the CBGB line. Since Loktionov was not brought back, players like Sislo and Josefon would still have been better options than the CBGB line while still allowing players like Clowe and Ruutu to play a top-nine role. If a team had this data, this is one of the ways it could influence situational tactics.
Okay, you’ve heard me rail on DeBoer’s deployment for long enough. Let’s take a look at how the Devils perform in the next phase of the game: the Transition Phase.
In Transition
I touched on a lot of these points in my transition article from the summer , and John did a fantastic job on Sunday going into detail on the Devils zone entry work with data from Corey Sznjader representing fifty-two games from this past season. For those of you unfamiliar with zone entries, I invite you to read Eric Tulsky’s paper from the Sloan Sports Analytics Conference at MIT from a few years ago. The main takeaway is that players entering the zone with possession lead to more shots for the team as opposed to the dump-and-chase style of play. Similar to John in his article, I’ll focus on the forwards in the neutral zone phase.
In his article, John identified the players that carry in versus those that dump in more often. Another way of thinking of a player generating a shot attempt in transition is to think of it as a player generating a carry in. Since a pass needs to be completed in order to generate the shot attempt, the recipient would need to enter the zone with possession, make sense? Take a look.
![Fwd_entry_assists]()
I’ve taken John’s chart from Sunday and will be adding a few columns to it. Here you can see the "Entry Assist" column, which means a player made a pass in the defensive or neutral zones that resulted in a shot attempt. When I tracked passes during the season, it had to maintain possession or result in a hockey move (shot attempt or pass), in order to count, so I was unintentionally tracking possession entry assists.
From the data John presented, we know Zubrus dumped the puck in an awful lot, but we also know that he generated the most entries. Now, was this more a product of playing on a line with Jagr and Zajac, who posted strong control rates on their entries? Most likely. But, it’s worth noting. From there, you can see down the list how many entries each player assisted on.
![Fwd_entry_contributions]()
Looking at the figures in total, we see Jagr on top by a healthy margin. Through his own carry-ins and entry assists, Jagr was involved in 213 possession entries, or 2.6 per game. However, looking purely at totals will often make players who played in more games look better than those that were more efficient. So, I decided to take into account each defensive and neutral zone start of the forwards.
Now, if a player attempts to enter the offensive zone, they most likely started their shift in the defensive or neutral zones. I totaled the number of zone starts by faceoff wins and losses from behindthenet and only looked at players who’d played at least 20 games with the club. Strangely, Loktionov’s numbers weren’t on the site, so I had to eliminate him from this study. Also, Ruutu’s data reflected his entire season totals between the Devils and Carolina Hurricanes, so I couldn’t parse out his data either.
![Fwd_entry_contributon_per_zone_start]()
The way to read this is simple: Starting with Boucher, of the 111 zone starts he received in the defensive and neutral zones, he directly contributed to a successful zone entry forty-nine times, or 44.1% of his overall zone starts in those two zones. So, almost half of the time Boucher started play in the defensive or neutral zones, he ensured the Devils entered the offensive zone with possession of the puck.
Now, it is a small sample size (only 111 zone starts) and some of those entries could have come from offensive zone starts, leaving the zone, and then returning with possession, but I went with what was more likely. It’s important to note the disparity between Boucher and Josefson, who had about the same sample size, but Boucher contributed to a possession entry almost three times as often as Josefson.
Moving on from Boucher, we see Jagr at 27.9%, Ryder at 26.5%, and Brunner at 25.9%, so it’s a mix of players that lead the team in entry contributions. At the bottom we see Gionta (surprise!), Carter, Josefson, Clowe, and Zajac. Zajac did have the highest number of zone starts in the defensive and neutral zones, so it was always going to be tough for him to post a high percentage. I think Loktionov would have looked good with these numbers, but, alas, the data was not available.
Onto the final phase: The Offensive Zone.
An Exit, an Entry, and Now What?
Hopefully a goal, but being Devil’s fans we’d settle for a legitimate shooting opportunity. Once the Devils finally enter the zone, or force an offensive zone face-off, how are they typically setting up their offense?
Above I’ve identified how DeBoer might benefit from using zone exits, zone entries, and shot generation data to, perhaps, better manage his situational deployment and usage of players in the defensive and neutral zones. In the offensive zone, I would venture a thought that less planning and thinking goes into who starts in the offensive zone: there’s less risk, so you’re not overly concerned with the two-way play of your players on the ice. You’re putting players out there with one job: get the puck and shoot it.
I also touched on this area of the ice in my piece on the offensive zone, but were I to put on a coach’s hat, I’d want the players that are most efficient at either taking shots or generating shots to start most often in the offensive zone. What you see below is a table illustrating how quickly a player either took a shot or generated a shot at even strength in the offensive zone (Even Strength Minute per Offensive Zone Shot Involvement, or ESM/OZ SI). Since we focused on shot attempts generated in the transition phase above, now we’ll focus on all shots occurring in the offensive zone.
![Def_oz_si]()
Looking again at the zone starts for defensemen, we see that Gelinas deserves to start in the offensive zone more than any other defensemen—not solely due to a perceived lack of defensive awareness, but simply because he’s involved in a shot on goal every 9:17 of even strength ice time, nearly a full minute faster than Zidlicky. Zidlicky also justifies his high zone starts as he’s involved in a shot every 10:11. From there, we see how awesome Andy Greene is: Greene is directly involved in a shot every 11:20, yet starts 12.7% more of his Zone Starts in the defensive zone than Zidlicky and 14.1% more than Gelinas. Fayne is just behind him at 11:29 per shot involvement.
Here we see Greene’s effectiveness in both defensive and offensive zones. Jon Merrill could be a boost in the defensive zone in exiting with possession, and the hope is that he receives a boost from Greene in the offensive zone, as he was next-to-last in shot involvement among defensemen. Oh, and the Captain still stinks. Salvador was involved in a shot 4:20 slower than Volchenkov. Let that sink in.
![Fwd_oz_si]()
Moving to the forwards we see how effective Mike Sislo was in the nearly two hours of even strength ice time he saw, being involved in a shot every 4:28, quicker than anyone on the team. Jaromir Jagr was right behind him at 4:29. Most of the more talented offensive players certain deserve to start in the offensive zone as much as possible, but Ryane Clowe actually saw more defensive zone starts than offensive. It works out as he was above-average in possession exits, and still managed to be involved in a shot faster than all forwards not named Sislo, Jagr, Patrik Elias, or Travis Zajac.
You would have liked to see Reid Boucher and Andrei Loktionov be more involved in shots given their high zone starts, but we’ve see Boucher’s contribution in neutral zone play as well as Loktionov’s stellar zone exits above. Again, knowing where your players excel and have limitations should be part of the discussion when deploying them on the ice. Looking at Josefson’s lackluster shot involvement, yet strong zone exit play, he’s a player, along with Loktionov, that I would have starting in the defensive zone more often. There’s nothing written anywhere that says defensive players have to be of the slow, gritty, and unskilled variety. Perhaps Loktionov and Josefson would have thrived in that type of role: get the puck and exit the zone with possession. We’ll never know, however, because Deboer reserved that job to players who are terrible at it.
Conclusions
What I’ve hoped to present here is a case to be made for all teams to track and house this type of data. Looking at exits, entries, and passing/shot generation statistics isolates which phases of the game players excel at and where their limits are. The coaching, or tactical, aspect of that is to take this information and use it to construct the best situational on-ice lineup you can. Defensive zone start? Give me a high PE%, low ESM/OZ SI player. Neutral zone start? Give me a high entry contribution percentage player. Offensive zone start? Give me a high ESM/OZ SI player.
Now, I’m aware a coach won’t always get the luxury of sending players into situations tailored for them, but the more often you put players in situations that play to their strengths, the smarter a coach you’re going to look.
Your Take
What do you have to say? Have I got this totally wrong? What do you think should fall under a coach’s tactical responsibility? Is this type of analysis what you’d like to see more of this season from us at ILWT? I will admit that this year I’ll be using more picture and (hopefully) video evidence in my articles. It helps to put a photo to the spreadsheet. Sound off below!
