When Sergei Bobrovsky took his show on the road and washed up in the Florida Panthers net, the writing was on the wall for Columbus. With no starter, they were toast this year. And Jarmo Kekäläinen decided to follow up his bold moves last year at the deadline with a bold lack of a move, and he just went with what he had.
What he had in net was a giant question mark. Actually, make that a whole series of question marks. During Bobrovsky’s peak years in Columbus, the backup chair was filled by Steve Mason for the first year and Curtis McElhinney for the next two and a half years. In 2015-2016, the Blue Jackets started to use their two prospects Joonas Korpisalo and Anton Forsberg, and they decided to waive McElhinney, which is how he ended up in Toronto.
Back in Columbus, they quickly chose Korpisalo over Forsberg and they let Forsberg go in the deal that got them Artemi Panarin. Meanwhile their 2014 third-round prospect Elvis Merzlikins was quietly working away in the Swiss league where all his good results needed an asterisk beside them that said: *It’s the Swiss league.
Elvis is now in the building
The question for Columbus fans was always: Would Elvis ever enter the building? And the answer came last summer when Korpisalo’s coronation as Bobrovsky’s heir to the starter’s chair (more or less by default and likely temporary) was interrupted by a usurper. Elvis was in the building, and he wanted that asterisk gone from his record.
The pair shared the net in Columbus this year, in what I think was not just a bold move by Kekäläinen, but a smart one. He knows what he’s got in ways most NHL GMs never do until a disaster forces them to find out at the wrong time. This Columbus tandem season has been described variously as one of the best in the NHL and a fluke of small samples for Merzlikins. No one ever agrees about goalies, but that’s quite a spread of opinion.
As we head into the last days of training camp, John Tortorella has consistently presented himself as undecided on who his starter will be for the playoffs, and he publicly defers to the goalie coach Manny Legace, which puts the Leafs into the position of not knowing who they’ll face.
My take, after a superficial look at their performance, is that the strength of the Columbus goalie tandem is that there’s two of them and they can largely replace each other right now. Korpisalo has the experience, Merzlikins has the recent success, and the Blue Jackets wouldn’t be in the playoffs without both of them. Rumours that Merzlikins is the next Jordan Binnington seem premature.
Who is in front of the net?
Before I take a less superficial look at these two Columbus goalies and see if I change my opinion any, we all need to understand the broad strokes of the team in front of them. If we’re not going to take that into account, we might as well just look at their All-Situations Save %, make up a narrative and move on.
At five-on-five, Columbus was 19th in score and venue adjusted Corsi For % with a dismal 48.87. (All numbers are from Evolving Hockey.) This is the part when you’re looking over a team’s results where you normally start expecting to see a bad defensive team that rode a goalie if all you have is that Corsi and Merzlikins’ Save %. However the Blue Jackets Expected Goals % has them ninth in the league at a very respectable 51.66%, and okay, I hate these decimal places. They’re silly and imply a fine level of distinction that just doesn’t exist. The Blue Jackets are at 52%, and that’s exactly the Expected Goals % the Maple Leafs produced along with Carolina, Colorado and Pittsburgh. These are good teams overall, but of that group only the Blue Jackets have a Corsi % below 50.
The Avs, the Leafs and the Hurricanes are all in the top five for Corsi For per 60 minutes with 60-61 shots per 60 (in this post, shots are all shots unless otherwise indicated). The Blue Jackets are 24th with 54. Six shots per 60, or about 5 per game depending on special teams time, doesn’t seem like a lot, and yet that’s the range between top offence and fairly dismal offence. They aren’t Red Wings bad (46 CF/60) or Sabres dull (50 CF/60), but they aren’t fast-paced.
Keeping with the Corsi theme, the Blue Jackets are only okay at shots against, allowing 56 per 60 minutes or one more than the Leafs (and we all know what they’re like). You might have been hearing about the amazing quality of the Columbus defence, and you might now be confused. But we’ve moved on from thinking shot rates are all there is to defending. When we weight all those shots by location and type and call them Expected Goals (which measures the chances of any one shot becoming a goal for or against) Columbus stops looking like a team only slightly better than Buffalo.
Well, in a way, they do.
Their Expected Goals For per 60 minutes is actually worse than their Corsi For. They’re 28th by this marker meaning they shoot so poorly, they actually erode the quality of their pitiful volume of offence down to something like what the, uh, checks this again, uh, Bruins produce. Columbus and the Bruins are tied in Expected Goals For per 60 minutes at 2.22. Other teams in this range are the Wild, who I remember as the most boring team in the NHL (again), the Devils and the Blues.
To this tepid offence, the Bruins add far superior ability to the Blue Jackets’ at limiting shots against, and when you add in their defending talent they get to first place in the NHL in Expected Goals Against at 1.98. The Blue Jackets are third at 2.08, and I’ll leave in the decimal places here where they mean a little. The Maple Leafs are a mediocre to bad-ish 2.45 by this measure and are in 18th place.
If you’re thinking the Blue Jackets are Boston-lite and you’re feeling a pit of doom open up, you aren’t totally wrong, but don’t overdo it. Boston has elite forwards and a power play that is also top of the charts. Columbus... does not. Columbus also start with a systemic disadvantage with their bad Corsi that they need to work hard to overcome. Columbus is very good defensively, but they have to first beat their own natures before they can beat the other team.
This is what that looks like:
They don’t merely defend well, they defend fanatically well in the area where the Maple Leafs are known to generate a lot of offence: the net front. Next time I’ll look at how that match up plays out before the shots ever get to the goalie. But for today, that’s what’s in front of the Columbus goalies, and this is what’s in front of Frederik Andersen and Jack Campbell:
Life is much easier in the Columbus net. And that’s the story save percentages on their own obscure. But we aren’t here to rank these goalies by value as if we’re handing out an award, we need to know if the Columbus duo is really going to hold up under a Leafs-level offence.
Back to the goalies
Moneypuck uses their own Expected Goals model to rate goalies, and when talk turns to just how the goalie has done, I like to use all-situations numbers because while for a skater, the special teams play is dramatically different in form from five-on-five, and we know that five-on-five success has predictive value, for goalies, stopping pucks is stopping pucks to some degree. Moneypuck’s model more aggressively weights the shots than EH’s model, and I find that shows up particularly on the power play.
By their model Elvis Merzlikins had the highest Expected Save % of any goalie who had at least 20 games played. In fact, it’s the third highest even if you include all goalies. Korpisalo is seventh. It’s easier to get to the top (or the bottom) of any list if you play less, but this model puts these two in a class with the Islanders goalies for how easy their lives were. (Expected Save % is what a league average goalie would do on that team.) Boston actually needed some goaltending this year, and their duo is down in the field closer to Andersen and Campbell.
Looking at Save % Above Expected — or measuring how much you’re better than a league average goalie — Tuukka Rask, tops the list with 0.823 percentage points, followed by Connor Hellebuyck at 0.774 (and he deserves the Vezina and the Hart). No one else is really in their class, but as you scroll the list, down, down, down, you pass the goalies who had a good season, even when disguised by bad teams like Mackenzie Blackwood and Jonathan Bernier (maybe the best play of his career), and you keep going into the negative numbers, past Mikko Koskinen and Aaron Dell, and just before Petr Mrazek, you come to Merzlikins at -0.124. Korpisalo is much worse at -0.375.
There’s good goalies with similar results because good goalies often have terrible years, but this is definitely in the starter having a poor yer, high-end backup area of results. Andersen and Campbell were both a little worse with their harder job and their uneven play (Campbell’s results include his LA numbers). They’re at -0.49, and they each have a different third digit, but they are essentially tied.
If I use Evolving Hockey’s different take on Expected Goals, but still stick with all-situations, Hellebuyck rockets to the top of Goals Saved Above Expected, and the order of the top players changes a little, but Merzlikins is 25th in the NHL (for goalies with more than 500 Fenwick Against, which is about 12 games) with 0.41 or a fraction of a goal over expected — making him league average. Korpisalo is way down at 47th and almost as bad as Frederik Andersen.
So even though this model approaches the question of how to measure shots and therefore goaltending differently from the first, it definitely does not think Merzlikins was one of the best goalies in the NHL this year.
Now, what do we do with all of that? Merzlikins isn’t going to be up against the full NHL averaged out over a season like that Columbus defence chart shows. He isn’t playing against whoever he faced in his 32 games played this year, his only NHL season. He’s playing the Leafs. And Andersen is playing only Columbus, and these two teams are each pitting their greatest strength against strength. Which is how everyone always says that, forgetting that means the weaknesses line up nicely too.
The goaltending for both teams is a weakness. One that mattered much more to the defensively average Leafs this season, while Columbus got good enough results out of one of their pair of prospects to help them make up for their truly terrible offence.
Merzlikins — and I’m just going to carry on assuming he’s the starter and this game-playing by Tortorella is his usual “look at me” monkeying around — has strength in front of him all game long to help him look better than he is. Anyone claiming he’s in Tuukka Rask’s class is insulting Rask, and I can’t believe I’m defending the honour of Rask, but here we are. Merzlinkins will be facing, not a league average offence, but the top-five Maple Leafs, and he has to hope that Columbus strength on defence is good enough to mix in with his good enough ability in net, and return the wins Columbus wants. The Leafs want the opposite, and what could go wrong facing a Latvian goalie?
Meanwhile Andersen has weakness in front of him that isn’t new to him, and that he has succeeded behind before, but he also isn’t facing a league average team. He’s facing a dismal offence that just seems to be crying out for some talent. Is he up to making up for the defenders in front of him? Or will he even need to, because if ever a team existed that could make the Leafs seem competent at defending it seems like it’s the Blue Jackets.
I did change my mind a little on these goalies. Korpisalo likely is just a decent backup and should be as irrelevant to this conversation as Campbell is — we hope. Merzlikins is only hazily understood with such a small sample of NHL games played, and Andersen, well, you know all about that. He’s had the worst year of his career. But goalies are not their last 5, 10 or even 30 games played, so how that all plays out in five rapid games over 9 days is anyone’s guess. If recent regular-season performance was all that mattered, we wouldn’t need to play the games.
It seems to me like the Leafs and Andersen will win the weakness war, and the strength war at five-on-five might be a draw. The difference could be how much the Leafs can keep down the Blue Jackets offence, and how well they can utilize their power play. Or that’s how I feel after a superficial look right now. But defending and generating offence is more than producing your averages, just like goaltending is. Next time I’ll see if a less superficial look at how these teams match up changes my mind.