Here we are, Leafs fans, at 20 games played, fifth in the standings (eighth by points %), and the early disaster-show of a team is all in the past. Twenty games is the magic moment when it’s okay to look at the team level stats. Prior to that they mean nothing, and anything could happen, now at 20 games, they are a hard and fast blueprint to the future.

Now, we all know that’s bullshit, but it is really easy to fall into the reductive trap of claiming smaller runs of game information is not just open to more variance, it’s actually meaningless. The corollary is that once you hit the magic 20, the information isn’t just predictive, it’s absolutely a perfect view of the team and the future.

All amounts of information that we collect about hockey are true and real: goals, save %, even secondary assists! Within the bounds of the NHL’s ability to gather accurate measures, they’re all truthful and real and describe games played. Some things, like shots and Expected Goals and other things calculated on the NHL data are more useful, but all of it is real. So if the Leafs put up a disaster show game, but it happens to be game six not game 26, it’s still them.

This is a tautology, but if they did that thing up there, that means they are capable of doing that thing up there. Which means if they did this thing, they could do it again, unless they become a better team:

So have they gotten better?

The ways in which they achieved that ignominious blowout loss to the Penguins do tell you things about the team, the coach and the players. The tricky question early on is: How often will that happen? The really tricky question is: Why does it happen at all, but that’s too much for a Monday look at the season to date.

Why do this after 20 games? It’s not magic, it’s just a factor of the extreme randomness of hockey results and the length of the season. When you have a lot of variance in a measurement — either random events or just unknown factors we can’t measure — the smaller the number of things you measure, the bigger the effect that variance can have. Once about 20 games have been played historically in NHL seasons (it’s about 23 or 24, actually) the outcomes observed in those games correlate the best with future outcomes.

Eventually, when there are a small number of games left, when the variance in the outcomes of the games yet to be played starts to have a big effect, the correlation fades away. You might ask if playoff games are a small number where variance has a big effect. Of course not! Everyone knows each playoff loss is a single individual’s fault for not trying hard enough. However, even in the playoffs, observed results are real, what the team actually did, and descriptive of their play. Hockey is not like getting a drop from a boss in a video game where if you fail, you just run the dungeon again. You only get the silver item if you actually play well a lot, over and over in all types of games.

On to the magic data:

The Good

For a while, the Leafs led the NHL in Expected Goals per 60 (all data is five-on-five and Score and Venue Adjusted, from Evolving Hockey unless otherwise noted). They’re now tied with the Panthers at 2.9. The Hurricanes are third at 2.88. Vegas, the Jets and the Flames are the only teams in the top 10 from the west. That means that pulling the hot offence trick is not unique in the east. There are only six western teams in the top 16 of the NHL.

The two astonishingly bad teams are Chicago and Arizona, with the Sabres, Canucks, Sharks and Kraken merely ordinarily bad. In other words, to find a weak offensive opponent to beat up on, you have to go west or to Buffalo. And watch out for Detroit, they are hot stuff offensively so far this year.

In some recent games the Leafs’ possession game, as measured indirectly by their Corsi For %, has been good, but eroded with poor quality shots to a worse Expected Goals %. This is not their norm, however, and their Corsi For per 60 is fourth, Fenwick is sixth, with that hot Expected Goals number even better.

Their overall percentages are also good to very good. The Corsi % is 53% and fifth, Fenwick % is 52% and 11th and Expected Goals is 55% and third. Twenty games might be magic but the Islanders game swung that last number by a full percentage point. One way to look at these numbers is that the Leafs’ quality weighting of their shots is just enough to make up for what they lose going from all shots (Corsi) to unblocked shots (Fenwick).

It’s worth a look at how that drop from Corsi to Fenwick is happening. The Leafs lose two ranking positions going from Corsi to Fenwick in shots for, so they are seeing some of their  shots blocked. They drop three ranking positions in the against side, which means they block a little less than the league on average, which is normal for the Leafs.

The Mediocre

There was a media talking point recently — and it is absolutely clear the Maple Leafs players and coach read from a script — that said their defensive game had tightened up with all five players on the same page. In one sense, that can be seen as valid in the overall percentages where it is clear they have the puck most of the time. The goal of defence, after all, is to get the puck and get up ice.

But it’s not at all true to say their shots against, by any measure, are better than past years. If the Leafs are consistent, it’s in their defensive mediocrity. They are not a disaster like they used to be. They aren’t showing decent percentages but making massive blunders due to lack of skill on the back end like they were in the Hainsey/Zaitsev/Polak days either. But I think we all know Muzzin and Holl are challenged this year as the pair with the most minutes and the most asked of them. A lot of bad results are coming when they are on the ice.

In Corsi Against per 60 minutes, the Leafs are 14th, Fenwick is 17th and Expected Goals are 15th. You can’t get more middle of the pack than that.

The Unsustainable

The measures most infused with variance, so much so, they can stay unsustainably high or low for very long stretches, are shooting percentage and save percentage. The Leafs are in the basement with a 6% Shooting percentage, good for 31st in the league. Thankfully, there’s 32 teams now, so they aren’t last. As discussed elsewhere, given the Leafs very good offence, it seems likely that number will rise. When and how much is a mystery.

PD Oh No: The Maple Leafs, Percentages and You

From a team Save % perspective, the goaltending has been good, but not spectacular. Their Save % is 11th overall and is likely very sustainable. But there is no guarantee, and you can’t count on that continuing all season. It could get better, worse, fluctuate all over the place, no one knows. Never bet on goalies.

In all-situations goals saved above expected from Moneypuck, Jack Campbell is second in the NHL, so he’s driving those good results. It behooves me to mention that Sergei Bobrovsky, widely considered a bust, is first and Frederik Andersen is third, so make of that what you will. Even Campbell’s NHL-leading 16 games played is likely too few to really draw too much from in terms of hope, despair, or any other emotion.

He made those saves, though, so he is capable of making those saves.

Déjà Vu All Over Again

Didn’t I write this article last year, the year before, the year before that and so on? Didn’t I used to make complex charts to really dig into the measurement of the Leafs? I did, and frankly, I could just take one of those and post it, you’d never know the difference. The Leafs have become better at offensive pace by a very small amount. When you’re near the top, there’s not much higher to go, but their defensive results have hovered at various rankings in the muddy middle of the NHL under Mike Babcock and Sheldon Keefe.

Their execution is better in the defensive zone, but the game against the Penguins tells a story. No not that disaster up there, this one:

The Penguins almost totally destroyed the Leafs offence. The Leafs don’t execute the same simple, yet effective, disciplined five-man defensive zone system. They allow more, so when their offence is stymied that’s what you see.

When their offence is allowed to roam free, you sometimes get this:

A little goalie trouble and the other team wins anyway. Or, you get a win like this when Campbell is on his game:

What this means is that the Leafs really are very unlikely to get a defensive win against a good offensive team. They just aren’t good enough to do more than meet the league on average on an even playing field. They have to have offensive power and good enough goaltending at the same time to win. When they have that, they are overwhelming. But they can’t let a simple defence shut them down because when that happens they don’t have a winning formula to pull out of their back pocket.

Same as it ever was.

Analysis