Do the New Jersey Devils lack depth?

CJ Turtoro tackles that question in his latest.

Follow @InfernalAccess on Twitter

Be sure to join the Discord channel to talk hockey, and everything else, with our writers and subscribers.

---

By CJ Turtoro (@CJTDevil)

As is always the case once we hit this time of year, the focus for the New Jersey Devils is on what players we need to acquire to fill perceived holes and inch closer to a return to the post-season.

I have some ideas about players that it would be nice to acquire through various means. Before we get into that, I think a sanity check is warranted on what our “holes” actually are.

One way to decide what to add is to pick a specific skill (example, PP trigger, PK defender, transition depth forward, etc,). And those can be evaluated on a case-by-case basis. I think an important starting point is just to have an objective understanding of where the Devils talent distribution stands relative to that of the rest of the league.

The most common way of doing this is to show a player’s value according to some metric (GAR, xGAR, etc.) relative to a typical NHL player in that role (say, 1st line center). However, I think that this binning can sometimes under- or over-estimate what a player’s value should be because not all 1Cs play the same amount of minutes. 

For example, Tage Thompson led Buffalo’s centers in both faceoffs and ATOI whereas Connor McDavid was 2nd on his team in both to Leon Draisaitl. So Thompson is being played as the 1C on his team at 17:53 per game, whereas McDavid is being played as the 2C at 22:04 a game. This doesn’t seem like an honest way of comparing their roles on the teams. A more reasonable way is to say “a team that gives X minutes to this player expects a Y level of production”. 

Instead of binning, I created a simple model to calculate the expected value per hour (in Evolving-Hockey’s xGAR) for a player based on the usage level that each Devils skater got (in terms of ATOI). Those metrics are, unsurprisingly, correlated – better players tend to get more minutes. Here is an example with a simple linear model attached to each position.

The formulas end up being something like xGAR ~ 0.05*ATOI - 0.6 for forwards and for defenders, it’s xGAR ~ 0.03*ATOI - 0.5. So, for forwards, you assume someone who gets 0:00 minutes a game would be 0.6 goals of value per hour worse than a replacement-level player and for every additional minute they get, the number goes up by 0.05 goals. 

This is a simple version of what is happening in the model, but I used a general additive model with a spline instead of the simple linear one you see above. All that is is fancy math speak for “I made the line bendier”. This offers a marginal, but statistically significant improvement in predicting the top-line guys.

Overall, this should do a better job of adjusting for the level of talent of both the player and the team around them. Here’s the results based on that evaluation. You can play with the Tableau here.

Forwards

For the Devils forwards, you can discern three or four distinct tiers. The top two centers (Nico, Jack), then the rest of the top-6 (Bratt, Zacha, Mercer, Yegor), and then the depth/fringe forwards.

In terms of those tiers, Jack grades out as an elite producer already and Nico is a league-average 19-minute forward. Combined, they are rather substantially above-average. In fact, they grade out as nearly identical to Kucherov and Point in terms of both value and usage. 

So our elite top two centers are pulling their weight just fine. The rest of the top-6 is a mixed bag. Jesper Bratt wildly exceeded his per-minute value, Zacha and Mercer significantly underachieved, and Yegor was about as expected. That makes them roughly league-average for players of their deployment as a group.

For the depth forwards, Tatar, Vesey, and Johnsson were all basically within the margin of error for players with their deployment. I think the Tatar one may surprise some people. It’s important to note that, regardless of expectations, he was treated as a bottom-six forward and produced basically what you’d expect from someone in that spot. 

This is an example of why you need to pay for scarcity. It’s why GMs pay top dollar to retain elite talent even if the end of the contract is concerning; exceeding value at the top end of the lineup is such a luxury and covers up so many holes down in the lineup.

For example, Tampa Bay had the one of the best teams in NHL history in 2019 and lost their 4th, 5th, and 6th highest scoring forwards (Miller, Johnson, Gourde) – the equivalent to an entire 2nd line – and they are now threatening to win their 3rd cup in a row since doing so.

Hughes and Bratt are way exceeding their expectations and Hischier is keeping up with his. Locking in those top three spots at above-expected production is indispensable. Make no mistake: Hughes, Hischier, and Bratt are absolutely capable of being the reliable core that Kucherov, Point, and Stamkos are to Tampa. 

Defenders

In terms of deployment, the defenders also fall into a few distinguishable tiers. We had a clear #1 defender in terms of TOI last year – although it was not who we expected it to be heading into the season. Severson played over two minutes a game more than the #2 option, Dougie Hamilton. Gravesy and Siegs were right there at comparable deployment to Dougie. Then we had a downward-sliding tier ranging from “regular bottom pair” (Subban) to “barely ever used” (Jaros) that we’ll call “depth D-men”. Then, there’s Mason Geertsen.

While there’s probably some disagreement among film scouts as to what defender(s) is/are responsible for the strong shot and chance prevention of this unit, the fact remains that our top-4 was above league-average.

If we shifted around the TOIs, Siegs would be a plus 1D, Seves a plus 2D, and Graves a plus 3D. Factor in the Dougie was playing through several injuries, and you have to think we enter the season with a perfectly usable top-4.

In terms of the defensive depth, when healthy, Subban overachieved and Smith underachieved. Between Subban’s strong play, limited minutes (which reduces his price tag), his winning of the King Clancy, and the Devils entire AHL defense failing to prove they can handle NHL work; whispers of interest in re-signing him have begun – something that was nigh-unthinkable coming into the year.

Meanwhile, Smith has done the exact opposite. There’s basically nothing good to say about Smith’s season except that it’s over. His rookie season was pretty substantially overrated and in my opinion he should enter 2023 expecting to need to earn back the roster spot he rightfully lost this season. 

The story of the defense is going to be one of maintenance rather than ascent. With Zacha’s (24) contract uncertain, the oldest returning member of the forward top-6 is Yegor Sharangovich who turned 24 earlier this month.

He is younger than the youngest member of the defensive top 4, Jonas Siegenthaler. Graves (26), Severson (27), and Dougie (28) are all on the wrong side of the aging curve. If none of the current AHL/NHL fringe (yes, I’m including Smith in that) are going to earn a spot, that puts added pressure on Luke Hughes to carry the future of this blueline on his shoulders.

Summary

So with those reviews, let’s now look at a summary of where every tier of each position group stands relative to where a comparable NHLer with that TOI would be expected to be. I’ve split it up into the aforementioned tiers – Stars (Jack/Nico), Top 6 Fs (Bratt/Pav/Yegor/Mercer), Depth/Fringe Fs, Top defender (Seves), Top 4 D-men (Graves/Dougie/Siegs). Depth D-men, and Mason Geertsen. 

The left column is how many goals added per hour they offer relative to a league-average set of players with that deployment (weighted for TOI). The right column is just converting that to how many minute it takes for them to add a goal.

For Nico/Jack, they add one goal over that of a typical pair of ~19-minute forwards every 367 minutes of ice time they play. These are non-overlapping groups so “Top-6 F” does not include “Team Leaders”.

What does this roster need in terms of skater tiers? Debatably, not much! Severson is a below average #1 defender, but I think that’s misleading because when everyone’s healthy he shouldn’t be there – Dougie should. If we plug him in, the rest of the top-4 is adding roughly a goal of value over an average team every 1000 minutes. 

The two lowest tiers other than that are still above-average in the rest of our top 6 and our bottom-pair D. I think they represent the most potential for growth because both really boil down to one replacement each – Ty Smith and Pavel Zacha .

I think Mercer’s analytics will catch up with his eye test next season and, with Zacha’s contract expiring, improving on him with a bona fide top line winger is really all we need to make this a formidable top-6.

Meanwhile, if a chain is only as strong as its weakest link, then our chain was blown to pieces every time Ty Smith was on the ice. The Devils were a -34 at 4800 minutes of even-strength play this season, and -23 of that happened in Ty Smith’s 1000 minutes.

If he’s not going to improve, then the Devils need to plug in someone to the 6D spot that will just be boring. If literally nothing happens when they’re on the ice, that’s a substantial improvement on the status quo.

Oh, and if Subban leaves, we need two of those guys. 

So after all this analysis, where do we land? Pretty much exactly where you probably were heading into this article. We need our top defender to play like one, and we should add another top winger. So, with that, I’m happy to have wasted your time and mine. Cheers y’all!