This trend is a major concern for TOR. They've never been known for playing tough defense but their trend of late has them as perhaps the NHL's worst. It's too much to ask of Andersen long-term and it's clearly too much to ask of any available back-up. pic.twitter.com/lfKubISgzG
— Sean Tierney (@ChartingHockey) December 5, 2019
I saw this, and couldn’t argue with the conclusion. So, I said to myself: Just how good a backup would you need on this team, not just lately when the defending is just horrible from the entire team, but overall this season?
The first thing you have to know to answer that is how many goals allowed would win each game. That’s easy to calculate, that’s the Goals For in each game minus one. This is real-world results analysis; this isn’t about how the team is doing five-on-five or what their luck is like or even if they’re good on the percentages.
Once you know the winning condition, it’s a simple thing to compare that to the All-Situations Expected Goals Against and see how many goals above or below that a goalie has to save to have won that game. Expected Goals Against is just the total goals that would have been allowed by a league average goalie on the number and quality of shots actually taken. I’m using Evolving Hockey’s Expected Goals throughout here.
This is the Leafs this season:
This is not a pretty picture to say the least. In any given game, a good goalie — a starter like Frederik Andersen — can and will save over two goals above expected. But not one of them averages that. And the vast majority of backup goaltenders are below league average. The Leafs have played nine games where a below league average goaltending performance would win it. That’s pretty normal for teams that are at least competitive. Both Carolina and Philadelphia have a similar number, but you can’t guess in advance which nine games those will be to put in your backup.
By this view of the game, the whole game, that debacle in Philadelphia really was the worst performance by the Leafs this season. The debacle against Pittsburgh barely registers, although it was bracketed by two other debacles, so it all seemed horribly unendurable.
Other teams have debacles too. The difference between them and the Leafs is that their run of the mill games mostly require close to one goal saved over expected or less. The Leafs run of the mill game is pushing closer to two goals saved over expected and three isn’t uncommon.
To answer the question of how good the Leafs backup would need to be, I simply took all the goalies in the NHL who have faced at least 50 unblocked shots and calculated their Goals Saved Above Expectation per 60 minutes. We’re talking about whole games here and imaginary regulation wins, so 60 minutes is per game.
To set your expectations of how many goals a really good goalie actually saves over expectation, look no further than the league-leading starter Conner Hellebuyck who has saved 10 (plus some fractions) goals over expected in his 22 games this season. And yes, that’s less than one-half a goal over expected on average per game. You can, if you look at that chart above, count how many times an average outing from Hellebuyck would have won the game.
If you look at the entire list of 69 goalies who have played this season, and what their results look like, you can, if you’re cruel, figure out how many of them, putting out his average performance level, could have won the game for the Leafs.
Do you see the problem? Even Frederik Andersen’s average outing isn’t good enough to win most of the time. No one’s is most of the time, which is true for all but the best teams. Sure, Freddie is above his own average somewhere around half of his starts, but even that’s not enough to get you a lot of wins when it’s always so hard. And for a backup? Forget it. Jaroslav Halak is good enough to win as often as Andersen, and so is whichever goalie you think is the backup for the Islanders. Carter Hart could handle it, just. Corey Crawford or Robin Lehner would work. And that’s about it. There’s a good number of starters in the NHL this season who could not provide value as a backup in Toronto.
No one the Leafs has put in the net as a backup the last two years has been good. And it doesn’t matter. Keep playing Michael Hutchinson, maybe against the weakest competition, but don’t spend on a backup until the winning condition for games is a feat doable by someone other than Andersen on his best night.
The question I answered here was this: Are the Leafs outshooting their problems like they need to, to get away with such low-quality defending? And the answer is an unqualified no. They aren’t outshooting their problems, and their problems have been getting worse.
Update: Cole Anderson posted a new look at the backups and starters using his lift over expected model. He shows the distribution of results for all the goalies. I added a line a little to the left of Andersen’s mean, and now you can see visually how few orange goalies (backups and lesser used tandem partners) have produced much in the Andersen’s average or better range, or the range of performance needed to get wins for the Leafs:
