2021-2022 New Jersey Devils analytics aWARds

CJ uses WAR models to see which Devils provided the most value across various aspects of the game.

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

---

By CJ Turtoro (@CJTDevil)

Let me first say that, before looking at this rather silly article, you should check out Todd’s actual year-end awards; which are more likely to be taken seriously.

This frivolous exercise is moreso meant to give people an idea of what the most prominent value model in the public sphere tends to see in players and from where it derives its final metric.

For those who don’t already know, WAR is an acronym for Wins Above Replacement, pioneered initially in baseball to describe a player's value to their team by determining how many runs they contributed beyond that of the typical player that would replace them. It’s since been exported to the NBA, NHL, NFL, and beyond.

WAR-on-ice was a hockey analytics website run by Andrew Thomas and Sam Ventura that was founded with the mission of calculating such a figure for hockey. They did and then bounced right into NHL jobs.

Emmanuel Perry’s website Corsica took up the mantle and he would eventually be hired by the Seattle Mariners.

Currently, the most frequently cited WAR model is by Luke and Josh Younggren and can be found on their website, Evolving-Hockey. They have two models: GAR (goals above replacement) and a model that mirrors the one originally created by WAR-on-ice, which they’ve dubbed xGAR (expected goals above replacement). 

The one we will use for the aWARds is the GAR model. The logic of model is based on the premise that each player has an influence on what game state the teams play in (penalties drawn and taken) and then an impact on how the play goes in each of those states.

There are offensive and defensive values in all 3 states (Evens, PP, and PK), but power play defense and shorthanded offense are too noisy so they are excluded. What we are left with is:

Even-Strength Offense

Even-Strength Defense

Powerplay Offense

Shorthanded Defense

Penalty Differential

The primary source for each of these awards will be GAR, but I will also cite xGAR if the picture is complicated.

Both total and rate (per 60) versions of GAR will be considered. So there is subjectivity involved. Without further ado, here are the 2021-22 NJD aWARds:

Even-Strength Offensive Value

We could go two directions with this and one is a LOT more controversial than the other. The uncontroversial one would be Jesper Bratt. In terms of xGAR, his 17.2 goals of value in this category (EVO) blow away the second-highest value which was Jack Hughes at 9.8. In terms of value per minute, he was one of the top 10 most valuable even-strength offensive producers in the NHL, ahead of even Connor McDavid.

That’s the obvious one.

The controversial one is: Andreas Johnsson. In terms of RAPM impact on 5v5 goals for, Johnsson has the highest per minute even strength value on the team. That is mirrored in his regular GAR value (EVO) of 10.7 which is tops on the team.

Upon inspection, though, this is largely fueled by an on-ice shooting percentage of 10.7% which leads all Devils. His past two years were 6.7% and 7.4% and he’s not generating dangerous passes in manually tracked data, so, there’s no reason to think this is a genuine skill. 

This is why WAR is informative, but not omnipotent. It would be accurate to say that knowing Johnsson was on the ice for a shift would make you more optimistic in that shifts even-strength offense than the shifts of any other player. But the reason for that confidence wasn’t necessarily Johnsson’s performance as much as dumb luck. Because of this, the winner is Jesper Bratt.