Photo Credit: Dan Hamilton-USA TODAY Sports

Examining William Nylander’s varying production rates

If you were to take a quick glance at William Nylander’s point production on a per game basis, you’d probably come to the conclusion that he’s performing more or less the same as he did last year.


He’s shooting and scoring slightly less, but producing assists at a slightly higher rate, which makes for a nearly identical points per game rate. When you dig a little deeper, though, and into his 5v5 rate statistics on a per minute basis, a very different picture begins to paint itself.

data via corsica.hockey

It turns out he’s attempting shots at a very similar rate to last year, but the rate at which they’re turning into shots on net has taken a decent hit. What’s really curious about that is that when you look at Nylander’s 5v5 shot map, his shots this year are significantly more concentrated in the dead centre of the slot. Most of that differential seems to reveal itself in the number of shots he’s having blocked this season, as he took 10.96 unblocked shot attempts per hour last year as opposed to his 9.63 per hour this year. He’s taking a lot less shots from the top of the faceoff circle than last year, which leads me to theorize that one of these seasons is a bit of an outlier in terms of the rate of shot attempts turning into shots on net. Was he getting lucky to have so many shots get through last year, or is he getting a little unlucky this year? It’s something worth keeping an eye on going forward.

Advertisement - Continue Reading Below
*shot maps via the wonderful hockeyviz.com, of course.

The improved shot location this year also shows in his individual expected goals, where his expected goal rate is only slightly lower than last year, despite the dip in shot rate. It’s also probably worth noting that he’s underachieved his expected goal rate for the second straight year, this year even more so than last. Based on the improvement in his shot quality I would expect some more positive regression moving forward, but, again, it’s worth keeping an eye on. It’s probably mostly luck, but as the sample size grows you have to consider the notion that it might be something to do with shooting talent. We know he has a great shot, but maybe there’s something to it. Maybe he’s trying a little too hard to pick that top corner too often because he’s perfect and it causes him to be a perfectionist. I don’t know, just throwing out some theories.

Though Nylander’s individual goal scoring rate has declined, his assist rate has exploded. His total 5v5 assist rate has nearly doubled and his primary assist rate has more than doubled. When looking at his shot assists and total shot contributions (thanks to data manually tracked by Corey Sznajder and visuals by CJ Turtoro) he’s actually contributed slightly less in the 20 games tracked this year than in the 28 games tracked last year, though still creating offence at an elite level.


There could be any number of explanations for this. Maybe a bunch of the games tracked this year were without Auston Matthews in the lineup, for instance, but whenever the underlying numbers look similar and the production rate explodes it’s usually helpful to look to on-ice shooting percentage and secondary assist rate for an explanation. As illustrated above, Nylander has seen an explosion in primary assist rate rather than secondary assist rate, which is a good sign, as primary assists have proven to be a repeatable skill whereas secondary assists are more often noise and tend to vary wildly from year to year for no real quantifiable reason.

Nylander’s on-ice shooting percentage, though, has just about doubled from last season (6.57% last season vs. 12.99% this season.) That probably explains a large portion of it. He probably deserved a few more assists last year and his assist totals might be a bit inflated this year, but it would be too convenient of a copout to simply shrug my shoulders and say that it should all be attributed to luck. So, to the on-ice shot density maps!

Advertisement - Continue Reading Below

you should really support Micah McCurdy’s hockeyviz.com, if you’re able.

It’s important to remember here that Nylander plays with the two guys who shoot almost exclusively from directly in front of the net in Matthews and Zach Hyman. With Nylander becoming increasingly similar to his linemates in his shot location tendencies this year, it shows in the heat map. It seems likely that improved shot quality has played a part in Nylander’s on-ice 5v5 shooting percentage explosion this year, but it’s hard to imagine Nylander and Matthews consistently being first and second in the entire NHL in 5v5 on-ice shooting percentage, as they currently are right now (among players with at least 500 minutes played.) All things considered, I think it likely that Nylander’s assist rate will fall somewhere in between this year’s and last year’s over the long run, provided that he continues creating at a similar rate. This is without taking into consideration that he’s still a couple of years from a typical player’s prime and will likely improve significantly in the meantime.

Nylander’s assist rate has lifted him from second line production on a per hour basis in both points and game score last year to being comfortably in first line production by both measures this season. To put it into context a little more, take a look at the three players just below Nylander in game score per hour this year versus last year.

*This is not meant to be a slight against Jonathan Marchessault. He is very good and has been even better this year. This is not meant to use Dom’s model against his own son.

’17/18 Nylander game score per hour comparables via corsica.hockey

*side note: seeing my wonderful sons Jeff and William so close makes my heart full.

’16/17 Nylander game score per hour comparables

The three names directly below Nylander in game score/60 this season are significantly more impressive than the three names below him last season. Again, no disrespect to Marchessault, who ranks fifth within these same parameters this season.

Advertisement - Continue Reading Below

So, how exactly is it that Nylander’s 5v5 production has increased so significantly over last season, but he’s merely on track to replicate his point totals?

It’s all in the PP production *winks so many times my eyelashes fall off and I hit my head off the wall, knocking myself out just as I deserve.*  As my friend/colleague here at theleafsnation.com Ian Tulloch lays out wonderfully over at The Athletic, the Matthews/Nylander powerplay unit can’t buy a goal at the moment. Last year, Nylander had 25 points on the powerplay (23 at 5v4 and two at 5v3) whereas this year he has only seven (six at 5v4 and one at 5v3.) When you look at the expected goals per hour with Nylander on the ice at 5v4 this year versus last year it looks pretty similar (7.63/60 in ’16/17 v. 7.38 in ’17/18) but, again, the heat maps suggest otherwise.

A quick glance at the shot density map would suggest the Nylander unit is simply creating less shots and it turns out that, surprise, that is the case. According to corsica.hockey, the Leafs attempted 104.63 shots per hour with Nylander on the ice last season and that number has dropped to 96.69 per hour this year. A lot of that seems to be correlated with the 13 games that Auston Matthews has been absent from the lineup this year, as during the 107:24 5v4 minutes played with both Nylander and Matthews on the ice this year they’ve created 99.4 shot attempts per hour. During the 23:30 5v4 minutes Nylander has played without Matthews this year, the team has created just 89.36 shot attempts per hour. Intuitively, this makes sense, as those two use each other almost exclusively on that unit and everybody else is just there to get one of them the puck and let them do their thing, more or less.

Ultimately, where Nylander has been very fortunate this year at 5v5, he’s been just as unfortunate at 5v4, which is basically the polar opposite of his ’16/17 season. It seems inevitable that the Leafs are going to lose James van Riemsdyk to free agency and, if they do, they should probably find somebody other than Matthews to take his spot on that top unit, because Nylander can’t make that second unit work by himself. Either that or move both of them to the first unit and play them 75% of the team’s total powerplay minutes.

Advertisement - Continue Reading Below

At 5v5, Nylander’s true talent level likely falls somewhere between last year’s deflated production and this year’s inflated production, but I’d wager it falls closer to the latter than the former.

All of these words and I haven’t even touched on what I believe to be Nylander’s strongest asset: his transition game, which I think will make him a top-tier play driving centre in the league, but that’s a story for another day.

This site uses Akismet to reduce spam. Learn how your comment data is processed.