Thatcher Demko was one of the best goaltenders in the NHL last season. He was a big reason why the Vancouver Canucks allowed the second-lowest rate of goals against at 5-on-5 and the Canucks led the entire NHL in 5-on-5 save percentage.
Demko was able to bail out the Canucks time after time last season, coming up with ludicrously improbable saves when things broke down defensively in front of him. He was so good that he made some people believe that the Canucks’ defence maybe wasn’t all that bad.
Goaltending can be volatile, however, with even the best goaltenders struggling at times. Depending so heavily on goaltending is a risky proposition but the Canucks did essentially nothing to address their defence in the offseason. They were counting on Demko to perform at his game-stealing best.
Demko has one of the worst save percentages in the NHL
So far, however, Demko hasn’t been able to steal any games. Through eight games, Demko has an ugly .876 save percentage — only Elvis Merzlikins has a lower save percentage among the 37 goaltenders with more than five starts.
As a result, Demko has given up at least three goals in all eight of his starts and four or more goals in six of those starts. It’s easy to think that if Demko had made a few more saves, the Canucks would have a few more wins this season and the alarm bells wouldn’t be going off in the Vancouver hockey market.
But is it really fair to put that on Demko? Anyone who has watched the games knows that the Canucks’ defence has been giving up a lot of Grade-A scoring chances through the first ten games of the season — is it reasonable to expect Demko to be stopping more of those chances?
There are a few ways to look at this. Let’s start with an expected goals model, like that of Natural Stat Trick. Expected goal models use the context of every shot — such as where the shot was taken, the type of shot, and whether it was a rebound — and assign each shot a probability of becoming a goal based on analyzing thousands of other shots from previous seasons.
At 5-on-5, Demko hasn't been that bad
According to Natural Stat Trick, Demko’s expected goals against at 5-on-5 this season based on the quality of shots he’s faced is 17.83. In other words, the average NHL goaltender would have given up approximately 18 goals against at 5-on-5 when facing the same shots.
Demko has given up 18 goals at 5-on-5.
In other words, Demko has given up exactly the same number of goals against at 5-on-5 as you would expect. Or, in other other words, Demko has been league average at 5-on-5.
Obviously, the Canucks were hoping for Demko to be well above average this season but this at least suggests that Demko hasn’t actually been bad — at least at 5-on-5.
There are reasons to believe, however, that Demko’s expected goals against doesn’t fully capture the quality of shots he’s faced.
These models are useful in many ways, particularly in large sample sizes over the course of a full season, but there are some things they simply can’t capture. The biggest is pre-shot puck movement — a pass or passes that mean the shot is coming from a significantly different spot than the puck was a moment earlier, making it a lot harder for a goaltender to track and stop.
An expected goals model can only capture that type of information inasmuch as other similar shots — shots from that area of the ice, with that shot type, and other similar contextual data — were also preceded by similar pre-shot puck movement.
The penalty kill numbers show why that might be misleading.
Demko’s expected goals against on the penalty kill this season, according to Natural Stat Trick, is 3.6. In other words, the average NHL goaltender would give up 3-4 goals when facing the same quality of shots.
Demko has given up 10 goals on the penalty kill.
But is that expected goals number really capturing how difficult those shots have been? More than any other situation in hockey, the likelihood of a goal on the power play depends heavily on pre-shot puck movement. Is Demko really giving up bad goals or are opposing teams getting scoring chances that are nigh-unstoppable?
In order to assess that, we need to look at the goals themselves.
The Canucks' penalty kill hasn't been good enough
I’m going to focus primarily on power play goals against but suffice it to say that there has been a lot of puck movement on the 5-on-5 goals, particularly on 2-on-1s where the defence hasn’t been able to prevent passes across.
We’ll skip over the first power play goal given up by Demko this season, as it came when Quinn Hughes was hit with an uncalled high stick, leaving the Canucks defending a sudden 5-on-3. Demko didn’t have much of a chance on that one.
Next up was this Connor McDavid goal on one of the slickest passing plays you’ll see on any power play this season.
The simple truth is that no NHL goaltender is stopping that puck. Demko has to respect that Leon Draisaitl could shoot from the bumper position and no one is expecting him to instead hit McDavid with that pass.
The next goal in that game was a shorthanded goal for the Oilers on a one-timer on a 3-on-1 that it wouldn’t be fair to expect Demko to stop.
If Demko had stopped either of those goals, it would have been held up as a save-of-the-year candidate.
The one power play goal scored by the Philadelphia Flyers in his next game should have been stopped, however.
Is the shot directly into the top corner? Yes. Were multiple people in front of the net looking to tip the puck? Yes. But still, it’s an unscreened wrist shot from the point — Demko had to have that one.
Likewise, you can blame Demko for this Alex Ovechkin goal when he faced the Washington Capitals, though you could suggest there’s a little bad luck involved.
While it’s a major red flag that the Canucks’ penalty kill gave up such a prime scoring chance to Ovechkin, there’s no denying that Demko accidentally kicked this one into his own goal.
The Capitals’ other power play goal in that game came on a deflection right in front of Demko that resulted in a rebound where no one boxed out Dylan Strome.
Heck, Demko might have made the save if Strome didn’t whiff on his initial shot, as he got his stick across to take away the bottom of the net. Is that goal Demko’s fault? I would lean towards no.
There was also nothing he could do about the game-winning goal, this one at 5-on-5.
That’s a complete defensive breakdown in front of Demko, who has to respect Ovechkin’s shot, leaving him vulnerable to the backdoor pass to Conor Sheary.
Let’s jump ahead to this power play goal by the Carolina Hurricanes’ Evgeny Svechnikov.
That is a massive passing lane wide open through the middle of the ice, letting the Hurricanes rocket a pass through the seam for a Svechnikov one-timer of which Demko can only get a piece. Is that goal Demko’s fault or the penalty killers’ fault for failing to take away the most dangerous passing lane on the ice?
The passing lanes on the penalty kill are again enormous on this goal by Nico Hischier of the New Jersey Devils.
This isn’t on Demko — the penalty kill cannot allow this type of passing. No goaltender in the NHL is going to stop that puck.
Then there’s this short-handed goal by the Devils later in the game, a 2-on-1 off a Vasily Podkolzin turnover that Oliver Ekman-Larsson played atrociously.
Later in the same game, another 2-on-1 where Tyler Myers may as well not even be on the ice.
There’s hanging Demko out to dry and then there’s hanging him on the clothesline in the middle of a hurricane.
Demko could be better but Canucks start is not his fault
It’s clear that Demko hasn’t been his usual dominant self. He has allowed some bad goals this season that are extremely atypical for him and he’s looked a little uncertain on stops that he normally would make with confidence.
Honestly, some of the above goals that seemed impossible to stop — Demko might have had a few of those last season. Demko himself would say that he expects himself to stop every single shot and he wouldn’t make any excuses for any of the pucks that got past him.
But there’s no way that Demko has been as bad as his .876 save percentage would suggest. Whether on the penalty kill or on odd-man rushes, the Canucks are giving up far too many cross-seam passes that force Demko to try to make impossible saves.
Until the Canucks’ defending improves, any blame aimed at Demko is misdirected.