How many goalies are out there are consistently "could also do a lot better"?
This is the same situation as Fleury back in the day. People would always say that Fleury wasn't good because he was just consistently average in the regular season and had major consistency issues. But at the end of the day "consistently average with flaws" tends to be a lot better than what most other teams get in net. You could get 1 year performances better than Jarry for less money, but you're very likely not getting consistently better performances for Jarry for less money.
Jarry is a microcosm of the issues with goaltending in the NHL as whole. "Consistently average with flaws" is better than what a ton of other starting goalies do in the NHL. Hell, Jarry was even better than guys like Vasilevskiy, Sorokin and Saros last year (in terms of GSAx) in this year that everyone says he was terrible. You can try to get better performances for cheaper, but the realistic outcome is that it's at best a coin flip for whether you'd be able to do that. Go ask Minnesota about how Gustavsson played, ask Toronto how Samsonov played, ask New Jersey how Vanecek played and so on.
Especially if we're talking about this season, who could you realistically do better with? Ullmark is the obvious guy, but it dries up super quickly after him. Maybe you can get a better 1 year performance out of Cam Talbot, but Cam Talbot in 2022-2023 put up an .898 save%.
There are a few problems with this.
First and foremost, it's leaning entirely on a singular projected metric as the defining "value" of a goaltender. And a very questionable statistic that can be wildly prone to team and system influences, with wild volatility year to year.
It's also leaning on an "average" statistic, compiled over the whole broader sample. ie. Designed to smooth out inconsistencies and nuance of how the guy played on any given night, specifically. It's statistically
not a measure of what you should expect to get from Jarry on any given individual start. It's a measure of what you should expect the ups and downs to average out to, over all starts. Those are two completely different, contextually critical pieces of information.
It doesn't really account for things like...hot streaks and prolonged slumps. Where say, a wild Ned comes in and starts to siphon off starts or "steal" his job.
The other aspect of this is just...yes, you're right in that it's risky and it's probably a "coin flip" if you can actually do as well or better with alternatives. Especially if you're trying to go cheaper at the same time. And it might even be a coin flip if you end up doing worse. But as i mentioned, it's an inherently "risk averse" strategy but also a little bit of a fallacious sense of security. It's investing "sure thing" money with major term in something that
isn't even actually a "sure thing". If you're investing $5.5M in something that you've got maybe a coinflip chance of replicating or exceeding for less money. That coinflip element cuts both ways. There's a 50/50 chance you're committing to worse goaltending.
Especially when...as an alternative, you might even get
two coin flips on your roster. Maybe even more if you can find some waiver exempt coin flips to stash (which imo, every team should be constantly drafting a steady pipeline of netminders to do).
But i mean...ultimately, if the whole thing comes down to GSAx...and it's telling you that Jarry is a better goaltender, or a better bet than Sorokin, Vasy, Saros, et al...i think it bears more scrutiny on what exactly that metric is telling you. I feel like a lot of the "goalies are voodoo" discourse comes from valuing things like that over actual performance.
Where goalies really are a bit "voodoo" or wildly volatile, is when you've got very mediocre "system goalies" guys like Vanacek that you put behind a wildly different team defence year over year, and surprise surprise...see wildly inconsistent results. Move a guy like Kuemper from a Cup Winner to the Caps. Or you've got a headcase like Samsonov who has always been wildly up and down and erratic in his actual play.
I can see why Dubas took the "risk averse analytics play". Not even sure it was the wrong thing to do for the Penguins given their situation. How much do you want to wager the last few years of "The Old Gang" on a coinflip in net? I think that was the reasoning. But as above...sometimes that coinflip still doesn't really work out that great. The Ned shot in the dark ended up probably better overall, especially at the cap value. And i'd be shocked if he doesn't completely fall apart if they decide to move Jarry and make him the starter. Because ultimately...not all coinflips are created equal in the NHL.