What We Learned: Can we really predict anything in the Stanley Cup Final?

So Game 1 of the most improbable Stanley Cup Final in quite some time is tonight.

I’ve seen a lot of people ask questions that are along the lines of, “How can ‘the numbers’ explain this meeting?” And plenty of pundits have also said, “It’s unfair to bring up future regression right now.”

Which leads one to wonder, naturally, what we could possibly expect to happen in the Final, because two teams that, by any reasonable stretch of the imagination, “shouldn’t” be here very much are here. And while you can’t delgitimize them by saying, “Well I’d like to see them run through their opponents again,” it is fair to question what we can expect from either team.

“ How do you explain Vegas ?!?” is a common refrain, it’s not hard to see this is the product of a great line and a just-okay goalie turning into the best at their jobs in the sport over a six-week stretch. Like I always say, people don’t want to hear that hockey is random, and perhaps getting a little more random all the time, but I gotta tell ya: Hockey is random. It’s not that “goalies are voodoo,” a common refrain in the analytics community, but that it’s impossible to predict when they’re going to get inexplicably hot.

Just about any goaltender can have 12 to 15 games of .940-plus hockey. It’s happened four times in the last six regular seasons, and just some of the names — Cam Talbot, Craig Anderson, Andrew Hammond, and Brian Elliott — don’t exactly scream “perennial Vezina candidate.”

The seven goalies who have done it in the playoffs since 2009 — Marc-Andre Fleury, Jonathan Quick, Braden Hotlby, Mike Smith, Jonas Hiller, Tim Thomas, and Tuukka Rask — are generally better (counterpoint: Mike Smith?) but there’s bound to be some survivorship bias in the postseason. Which is to say goalies with lower save percentages don’t often make it through 15 games, right?

So it’s really not that uncommon for a team that has, say, a 49 percent chance to win any given game to win 12 games out of 15 tries (ballparking the number it looks like about 1 in 8). Especially if their goalie gets inexplicably hot. That’s not something that reflects well on Computer Boys, who can throw their hands up and say “What can ya do?” whenever there’s something that well and truly defies the odds like a Marc-Andre Fleury hot streak of this nature, but if there’s no way to predict outsize performance, “What can ya do?” is the only thing you can say.

The question, then, is does he keep it up? I mean, some of those goalies did it — Quick and Thomas both won Cups from that group — but most did not. And if Fleury suddenly becomes a regular, normal Fleury again, posting something between the mid-.910s and the low .920s, who will be surprised? And if he keeps it up, who will have a right to be shocked?

Likewise, Washington. Everyone they played in these playoffs was a better possession team in the regular season. Washington (marginally) turned that around in the playoffs but is still getting a little bit out-attempted at 5-on-5, largely due to score effects.

Their playoff downfall has so often been inexplicable rotten luck when it came to getting the puck past goaltenders. No matter how well they seemed to play, their shooting percentages typically tanked, maybe due in part to the fact that playoff teams tend to be better at killing penalties, and also refs are far less likely to call postseason penalties in the first place. As discussed in the most recent Trending Topics , they’re getting the same high-quality goaltending as always but their shooting percentage is much more in-line with what they did in the regular season (10.5 percent now versus 10.8 percent then, both of which are quite high).

Is that sustainable in a series that will run between four and seven gams? I don’t know. No one does, really. So much of both teams’ success has been predicated on unsustainable runs at either end of the ice that you have to to say this is getting to be like the main event of WrestleMania III: the irresistible force (the Caps’ shooting luck so far) versus the immovable object (Fleury). Something has to give.

As you might expect, these are the teams with the two largest differences between expected and actual goals per 60 minutes in all situations. Vegas is plus-0.94 over what they “should” score every hour. Washington is plus-0.62. Only one other team, Boston, is north of plus-0.5, which I guess you can argue presaged their early bounce-out.

But now we’re pretty far beyond normal probabilities, right? Here, too, you’d probably have to give Washington the slight edge just because you can’t really count on Fleury to keep delivering you one-goal wins through .947 save percentages. But it could happen! And if it happens, there’s no way you can say, “I’d like to see them do THAT again,” because of course they could not. You win or lose the games you play and that’s all you can do. To ask for re-litigation on every bounce of the puck is idiotic, but that doesn’t mean both teams aren’t supremely lucky to have gotten this far.

In every sport, over a long enough timeline, the best teams win the most games and that’s true in hockey as well. Pull up a list of the best regular-season records over any three-year period since the NHL got advanced stats and the top five in winning percentage are going to look more or less like the top five in 5-on-5 CF%. That’s another thing we don’t need to re-litigate.

But to really get under the hood and look for reasons beyond randomness that Vegas and Washington got here? Like we all know Marc-Andre Fleury is why Vegas got here. We all know the Caps keeping up their hot shooting for once is why Washington got here. What else is there to investigate?

You want numbers to explain this? It’s not hard. Most hockey games are very slightly weighted coin flips, and it’s like Smooth Jimmy Apollo says: “When you’re right 52 percent of the time, you’re wrong 48 percent of the time.”

People want to ascribe deeper meaning to it than that — they want to figure out the “why” of it — and in the playoffs especially, you really can’t.

What We Learned: Playoff edition

Vegas Golden Knights: I know a lot of people have talked about how long of a layoff this has been for Vegas and how that might affect them. I’m kinda dubious that it will at all, because what’s the tangible difference in “rust” you’re going to develop over nine days versus the Capitals’ five? If they went right into the Cup Final — say, on Saturday after the Caps won on Wednesday — that might have been a thing, but if it’s nine days off versus five, I’m less inclined to believe in it.

Washington Capitals: The Caps Are Good On The Road. That’s the prevailing thought on them in this postseason. Which is interesting because their underlying numbers at home are very good, and on the road they’re a little underwater. You’ll never guess the difference: An all-situations home PDO of 98.1 in these playoffs, versus a road PDO of **drumroll** 102.7. Does that affect Vegas, does the Vegas Flu come into play, etc. etc. etc.? I mean, maybe. But also: Maybe not???

Play of the Weekend

There were no plays this weekend so here’s that Ovechkin one-timer from Game 7 again. Pretty good.

Gold Star Award

Here’s a fun stat: Vegas and Washington have both committed more penalties than they’ve drawn in these playoffs. Doesn’t seem like that should be a thing but here we are.

Minus of the Weekend

This John Vanbiesbrouck hire, I dunno. I have heard a lot lately about how hockey is for everyone, but then a guy who repeatedly called a black player maybe the worst possible racial slur — behind his back no less — gets one of the top “growing the game” jobs in the sport, and USA Hockey doesn’t even check in with the guy who was the target of his abuse. Folks, it’s pretty bad.

Perfect HFBoards Trade Proposal of the Week

User “ sabrescupbound” is doing a classic mix-em-up”

Signoff

Aurora borealis?

Ryan Lambert is a Puck Daddy columnist. His email is here and his Twitter is here.
(All stats via Corsica unless otherwise noted.)