After two games in Ottawa, it’s time to ask what we’ve learned about the Leafs. The top three forward lines seem as poorly optimized now as they did before opening night ever happened. I talked about that after the Montreal game, and the Ottawa games showed some of the same problems, but also showed some lines working better.

The NHL is the show not the rehearsal

The third line was dramatically better against Ottawa, and they dominated every matchup in the Friday game except against the Stützle - Stepan - Dadonov line. They totally shut down Brady Tkachuk and Josh Norris, though. On Saturday, all the Leafs dominated in shots (all shots, not shots on goal) over the Sens, but the Kerfoot line was the best at 82% Corsi. They only struggled against the Stepan line again, and that to a much lesser degree.

Most numerical stats are from Natural Stat Trick and are Score and Venue Adjusted five-on-five unless otherwise noted.

The theme of both games was poor shooting location. Friday’s game had the Leafs at 49% Corsi for and 41% Expected Goals for.  In other words, they shot the puck almost exactly the same as the Sens (the real difference is 2 shots less for the Leafs), but they were shooting from so far out, that eroded down to a situation where without the power play, they weren’t in the game.

Saturday was a total domination in Corsi, with the Leafs at 75%, leaving the Senators to try to subsist on only 22 shots all game, and yet the score was 3-2 and the Expected Goals was only 62% for the Leafs. The Leafs had 56 shots and only 36 unblocked shots, which whittled down to 26 when you get to shots on goal. Or, exactly what the Blue Jackets did to them in the playoffs.

The Saturday game was very evenly split by most xG models at all-situations. It was a domination in zone time, but the use of that zone time left the Leafs shooting from the point and the Senators right up in Jack Campbell’s face:

Maybe he should be lauded for his .895 save percentage. That’s as bad as anything faced by Frederik Andersen. On the Leafs side, this all-situations view makes them look better than they were because their power play really looked good, and resulted in some quality chances.

Assuming this is the intended outcome or a coaching philosophy is likely not wise, but it’s also not a good way to score a lot of goals. These types of passes are mostly the sort least likely to add value to the shot that comes after them.

Four points in three games in a perfectly fine start to the season, even if Montreal walked away with one and the Senators two, but aside from proving that the Kerfoot line can shut down almost everyone on the Senators, there’s really only one other bright spot.

Auston Matthews is shooting from everywhere, including such good locations, that he’s got double the individual expected goals of his nearest teammate (by a quick check on Evolving Hockey, not the above HockeyViz model). John Tavares is not bad, mind you, and his power play performance has been stellar, but Matthews is the team offence at five-on-five.

So far the Joe Thornton experiment is at least not nerfing the most important player on the team.

But that leaves a big question mark over the head of John Tavares and his linemates. They aren’t getting all that much ice time, in three games Tavares is 8.5 minutes behind Matthews in five-on-five TOI, and his line aren’t doing great things with the time they do have. They’ve been blessed with some points, but three-game points rates are so deeply soaked in randomness as to mean less than nothing. Don’t even look, lest you become less informed for knowing them.

This is not good:

It’s also not typical for his career. Anyone who has watched the games has not seen much from William Nylander at even strength, and Jimmy Vesey has barely registered as there. If the Thornton trick is at least one you can fake your way through for most of a game (he plays less than Matthews every game) the Vesey on the Tavares wing concept doesn’t seem to be cooking with any fuel.

This roster isn’t an easy puzzle to solve, and the idea seems to have been to play Wayne Simmonds up on that Tavares line a little to see how that would go. It hasn’t happened a lot, but it worked in a few shifts here or there.

So the dilemma is this: if the Kerfoot line is good shutting down some team better than Ottawa, but maybe not the Habs, and the Tavares line needs something to change, how do you fix this? Maybe this whole roster thing needs to be turned around and looked at differently.

It’s not about Sheldon Keefe stubbornly sticking with the (clearly) absurd idea that Joe Thornton can play top line, but rather, he’s trying to find the least bad alternative from Thornton, Vesey and Simmonds. He can’t replace Thornton and Vesey both with Zach Hyman. So unless Simmonds really shows more oomph  than he has so far, as he gets conditioned to playing again, one of Thornton and Vesey is staying in the top six. If that one is Thornton, the guy who spells him so he plays three minutes a game less than Matthews and Marner is usually Hyman.

This is not the first time my conclusion is that the Leafs need to clone Zach Hyman.

I’m a little less aghast at the lineup choices right now, not because the Leafs beat the worst team in the division in a way that carried the illusion of dominance, but because if Keefe is looking for the least worst of some bad options, that’s a mode of thinking I can understand. Maybe I’m rationalizing it, but it’s not like there’s a stable of top six wingers he just won’t play.

As the man on TV likes to say: We’ll see where this goes.