Upon a cursory glance, the Devils powerplay seems to be a thoroughly unnoteworthy aspect of the team. The Devils were 10th in PP% which is more or less what one would expect given the were 14th in total scoring. When looking into the advanced stats, they seem totally unremarkable as well. According to Corsica, the Devils on the PP were also 11th in GF/60, 14th in GF%, 20th in xGF/60, and 23rd in xGF%. That’s almost exactly league average production. So why is the Devils PP worth an article at this point of the offseason? Because it is the most obvious example of the Jekyll and Hyde that was the Devils team season. The constantly terrifying Mr. Hyde was the Devils with Taylor Hall, and the utterly benign Dr. Jekyll was the Devils sans-Hall.

The most visually striking way to get this conversation started off is with the WOWY charts from Dr. McCurdy’s Hockeyviz:

Those shot maps are stark. They are reminiscent of the Capitals with and without Ovechkin. As impressive as that pair of graphs is, it isn’t the best way to prove the impact that Hall specifically has. Those graphs just show with and without Hall, but that would normally mean with and without the entire PP1 unit. Any team with a really good PP1 unit might see something similar or even worse — Phil Kessel and Pittsburgh’s PP1 unit are a good example. So, the Devils have a worse PP2 unit than PP1 unit. So What? The problem is that it’s actually deeper than that.

The shot maps above are basically the equivalent to “Rel” stats. How good a team is with a player on the ice (left) vs. off the ice (right). But Rel stats can overstate the effect of a specific player if they spend a lot of time with a specific unit, as is prone to happen on special teams. A way around this is using “RelT” stats. These are not relative to the team, but relative to specific teammates. It’s essentially a weighted average of a player’s entire WOWY chart. The result is the net effect the player had on each teammate. If you want a more full demonstration of this method, check out EvolvingWild’s 2-part piece at HockeyGraphs.

When looking at RelT GF/60 (a player’s “goal impact”), Taylor Hall was a +9.39. This can me interpreted as saying that his impact was worth 9.39 goals per 60 minutes (~0.3 goals per powerplay). That figure was 3rd in the NHL among players with 50+ minutes, behind only Claude Giroux (+9.89) and James van Riemsdyk (+9.5). But Giroux played 96% of his PP time with Voracek and Gostisbehere who are also very good, and JvR was tied for 3rd in PP scoring on his own team. Hall was not glued to anyone on the PP1 unit and led the team in PP points by 14. I think it’s fair to say that he was the most valuable player to his team on the powerplay — unsurprising, I suppose, since he was the most valuable player overall.

Some people may be thinking at this point that I’m doing a disservice to Will Butcher who, as a rookie, was tied for 10th in the NHL in powerplay assists among defenders. First of all, I think it’s worth noting that “elite for a rookie” is different from “elite.” Butcher was very good and put together one of the most prolific seasons for a rookie defender in the dead puck era and one of the best seasons for a Devils rookie ever. But Butcher has Hall to thank for a LOT of his success. This is what the team looks like with each Hall and Butcher on and off the ice:

These were created using the Line Stats tool from Natural Stat Trick. If it’s not clear from the chart, let me put it into words. In 157 minutes with Hall, Butcher saw 32 goals for and 3 goals against on the powerplay. In 57 minutes without Hall, they scored 1 and allowed 1. In FIFTY-SEVEN MINUTES without Hall, Butcher was EVEN. So I ask you (*Marisa Tomei voice*), who driving the bus on the powerplay?

All of these stats are mostly implicit — they attempt to infer Hall’s effect by what happens around him. So let’s just put a cherry on top with the what Hall is directly responsible for. According to Corsica, Hall put up 36 points on the powerplay. The Devils scored 53 goals, so Hall was responsible for 68% of them — the highest mark of any player in the NHL (Seth Jones, CBJ D, was 2nd with 65%). He scored 9.9 points per 60 minutes — over a half a point higher than Blake Wheeler who was 2nd and over a full point higher than 3rd place Morgan Reilly. And it’s not just the direct points either — Halls GS/60 (Game score per 60) of 7.76 was also a league high.

This is going to need to be a point of attention for the coaching staff in the offseason. Departing coach Geoff Ward was never able to fix it, and the natural heir to his role, Ryane Clowe also has left to head coach Toronto’s ECHL-affiliate. Hynes, or whoever comes in to replace Ward can bank on Hall for production again, but will likely aim to get something more out of virtually everyone else than was given this year. This likely starts with getting the 2nd best player on the team, Nico Hischier, going. Nico struggled mightily on with the man advantage, scoring just one goal and one primary assist in 167 minutes (?!?!?!). The “Devils without Hall” are not a great team overall, and the powerplay, far from an exception, seems to be especially vulnerable.

Under Ward last year, as evidenced by the point percentage numbers above, the plan with Hall on the ice seemed to be “get the puck to Hall” and the plan without Hall seemed to be “wait until Hall is back on the ice so we can get the puck to him.” Unlike the MVP debate, where most seem to believe McDavid was the “best” player but not most valuable, I believe we can make a strong argument that Taylor Hall was both the most valuable AND the best overall powerplay skater in the NHL.

And the Devils were absolutely lost without him.

Your Thoughts

Is this more or less what you expected? Are you surprised how good he was on a league-wide scale? Or how bad we were without him? Do you think next year will be improved or more of the same?

Thanks, as always, for reading, and leave your thoughts below!

https://www.allaboutthejersey.com/2018/7/11/17550534/taylor-hall-is-the-entire-devils-powerplay

Leave comment

Your email address will not be published. Required fields are marked with *.