Usage is a tricky subject in hockey. For many years, maybe forever, the way a coach uses a player has been taken as proof of who the player is, how they played, why they scored or didn’t score — the whole narrative stew of a goal-based view of a game. We’ve invented words like “sheltered” because playing good offensive hockey and having the most value scoring goals is suspect. We’ve decided the trick of dancing backwards in skates that is defensive zone play is really done by slamming into someone at high speed, and that a defensive zone breakout is the hardest thing anyone has ever done in the history of sport, so only the toughest of players can start a shift in the defensive zone.

And then when the large pools of data get analyzed and we discover that zone starts have very little impact on a player’s results, that quality of competition is very even for most players over a season and teammates matter most, we tend to ignore all of this information too much, and the sport of looking at single-game on-ice shots for and against (all shots, not shots on goal) is born.

Single game results mean nothing out of context, while on a full season, the context of teammates overwhelms everything else. This is the basic shot maps for the game:

If you go to Evolving Hockey and look at the xG plots, you can see that the effectiveness of the Canadiens’ play fades when you weight for shot location, which is hardly news.

For the first game of the season, some usage trivia that might tell us a few things about the Leafs, the Canadiens and Sheldon Keefe’s choices that don’t show up on those charts:

Ice Time

note: the word shot means all shots, and all information is five-on-five

William Nylander played 1:37 with John Tavares and Mitch Marner, out of 2:40 total with Tavares. That amounts to about four shifts away from Alexander Kerfoot. He scored his goal with a secondary assist going to Marner (already on the bench) and Tavares on the ice.

Tavares and Nylander had the most minutes of all forwards, with Mitch Marner third with 13:48 to Nylander’s 14:16.

All six defencemen played more than all forwards except Tavares, with the third pair logging very close to 15 minutes in the game. The top four ranged from 16 minutes to 17.

The fourth line played just over nine minutes.

Ondřej Kaše played the least of any top nine forward at only 11:34. We should watch this as the season goes on, because he was so obviously very good, that amounts to being underplayed, so Keefe is being very cautious with his minutes.

The top three lines in their standard arrangement (so leaving aside Nylander’s shifts with other people) rotated out at almost identical amounts ranging from 10:13 to 10:58. The Maple Leafs without Auston Matthews are a more balanced team, which shouldn’t be news.


The defence pairs stayed very static with only tiny scraps of time on shift changes spent with some other partner. The forward lines were also mostly static beyond the creative use of Nylander.

Morgan Rielly played with David Kämpf, Michael Amadio, Tavares and then Kerfoot in that order, but the amounts of time range from 5:27 with Kämpf to 4:04 with Kerfoot.

Rasmus Sandin played with Kerfoot (6:07), then Tavares (4:45), Amadio (3:00) and Kämpf (1:21).

Jake Muzzin played with Tavares (6:03), Kämpf (5:58), Kerfoot (2:31) and Amadio (1:55).

So the plan was to use Muzzin and Justin Holl with the new “shutdown line” and never with the offensively focused second line. They were also tasked with backing up the top line. This can be seen as matching the defence to the forwards or as matching them to the opposition. Or both arising out of the same need — to put the defenders that can handle offensive pressure from the top competition on the ice with as much of it as possible.

What started out looking like rolling out three pairs in equal portions, ends up as extremely skewed usage when you look at the teammates.


Tavares v Josh Anderson was the theme of the forward matching. They faced each other for 5:49, as Tavares played both the Dvorak and Suzuki lines for over five minutes each. The rest of his time on the ice was spent rolling over the hapless Jake Evans for under four minutes and playing a couple of shifts vs the Candiens’ fourth line.

Kerfoot got the other five minutes of Suzuki time, but barely saw Dvorak (Anderson’s centre). He got to roll over Evans for 4:12, and barely saw the fourth line of the Canadiens.

Kämpf played 5:37 against Dvorak, 4:07 against Evans (the only time Evans wasn’t obliterated) and barely saw Suzuki or the fourth line.

Amadio played most against Dvorak — 3:07, and then almost three more minutes against the fourth line and three against the other two put together.

The plan seems to have been to shutdown Dvorak — and by extension the annoying Anderson — while leaving the rest of it to largely sort itself out. It’s extremely odd, therefore, that the Leafs fourth line played that scary beast of a line the most. They look bad by raw stats, and well they should. Shutting down top lines is not Jason Spezza’s or Wayne Simmond’s job anymore.

Where this matching become even more skewed is with the defence.

Muzzin saw Dvorak for almost nine minutes, and Dvorak only played 15.  Rielly was matched to Suzuki, and Sandin looks so amazing, largely because he feasted on that Evans line and only saw Dvorak for 1:31.

Now that skewed usage of the defence by teammates becomes even more skewed when we look at competition.

If you go back up to the shot chart and cluck your tongue over Justin Holl (because it’s always his fault) and the fourth line, you’re missing a lot of context.

On the other side of the ledger, the exciting results for the Kerfoot line are a little less glossy.

If you’re a Canadiens fan, you really should worry about that bottom six. How did the best depth in hockey turn into that, anyway? Only Claude Julien knows. At the same time, toss off a thank you to the Canes for forcing that trade for Dvorak on you. That’s working out well so far.

Numerical data is from Natural Stat Trick.