NHL Playoffs: Using analytics to see how the Conference Finals will shake out
The second round confounded both old school pundits and fancy stats analysts alike, particularly in the Eastern Conference, where the Penguins managed to once again dispatch the Capitals despite generating poor puck possession stats, and the Senators continued their Cinderella story run to their first conference final in a decade. As it turns out, our model got both of those matchups wrong.
The Western Conference was kinder to us in that we picked both Anaheim and Nashville, but needless to say it's hard to imagine many people predicted this final four last month.
So who does our model pick to go to the finals from here?
Penguins (111 points) vs. Senators (98 points)
Prediction: Penguins (70.8%)
The Senators’ improbable playoff run continues as they roll into the Eastern Conference Finals, and no surprise, our model has them as huge underdogs again. How have they managed to get this far, and can they keep it up? An oversimplified answer to the first part of this question can be summed up in a word: Erik Karlsson, one of the greatest defensemen to ever play the game. (Okay, that wasn’t one word, but maybe it should be.)
So far in these playoffs the Senators as a team have been middling in many dimensions. They have taken just over half (51.8%) of all shot attempts in their games and almost exactly half (50.3%) of unblocked shot attempts. They have had considerably less than half of the scoring chances (42.8%) and less than half of the expected goals (46.7%). Their power play hasn’t been particularly effective at 14.6%, either. The Ottawa penalty kill has been very good, however, the best of all remaining teams, at 87.5%.
God bless Erik Karlsson: The Senators think their defenseman is Heaven-sent
What is truly remarkable is the difference between how the Senators perform with Karlsson on the ice, and how they perform without him. With him, the Sens generate a whopping 57.1% of shot attempts compared to only 48.1% without him. And the shot attempts with Karlsson are also of higher quality, as Ottawa generates 53.3% of the scoring chances and 54.5% of the expected goals when he is on the ice, as opposed to only 36.3% of scoring chances and 41.4% of expected goals when he is off the ice.
While Pittsburgh is not exactly the juggernaut it was last year (they are currently last among all playoff teams at 42.1% on shot attempts, 49.4% of scoring chances for, and 45.7% of expected goals for), they do have one thing the Bruins and Rangers didn’t have: depth. As such, they could be a much tougher matchup for the Sens, unless Ottawa coach Guy Boucher figures out a way to get Karlsson on the ice for the entire 60 minutes (or, as has often been the case in the Senators’ playoff games, even more).
Our model doesn’t like the Senators’ chances of making it to the Finals, but they’ve bucked the odds so far. If they do make it, you have to think that Karlsson is the leading candidate for the Conn Smythe.
Ducks (105 points) vs. Predators (94 points)
Prediction: Ducks (60.0%)
Much like the Senators, the Predators didn’t get the memo that they weren’t supposed to be playing this week. To be fair, in an otherwise brutal year for our model, the Preds have been a bright spot, as we predicted their last two upsets. Unfortunately, this is where we part ways and think they’ll bow out.
It will pain some Maple Leafs fans (though they’ve perhaps finally moved on now that playoff hockey should be in Toronto’s foreseeable future) and cause many analytics types to proclaim that even a blind squirrel finds the occasional nut, but Randy Carlyle’s Ducks have become a solid possession team, generating 53.6% of all shot attempts, good for fourth-best in the playoffs and tops among the remaining teams. Nashville, meanwhile, has been fairly middling at 50.3%.
The Ducks under Carlyle have continued their habit of getting shot attempts on net, with 55.1% of all shots. As we noted last year, one measure of whether a team is just firing pucks out of blind desperation or actually shooting at something is whether the shot attempts actually get to the net. So far the Ducks are doing very well in this regard.
Pekka Rinne has been the story for Nashville, and we’re sure that people who follow our column will scoff at the punching bag we made of him during last year’s playoffs. But the strategy in Nashville is as it was then, which is to protect the goalie.
Nashville’s .955 save percentage during 5-on-5 play is simply a silly number and there’s no doubt Rinne is having a terrific playoffs. But there’s also no doubt that P.K. Subban, Roman Josi, Mattias Ekholm and Ryan Ellis constitute an embarrassment of riches on the Preds’ blue line. So it can’t come as a shock that Nashville has given up only 6.2 scoring chances against per 60 minutes, which is the best among the four remaining teams.
Anaheim meanwhile, has so far been the top offensive team this postseason, with 10 scoring chances for.
Ryan Ellis is lifting the Predators in the postseason as his own game rises
It’s hard to imagine there are too many serious proponents of the “offense sells tickets, defense wins championships” school of thought in today’s NHL, but if the Preds win, we will no doubt hear some pundit dust off that platitude.
To be fair, a defense as good as Nashville’s is certainly capable of beating anyone, but so is an offense as talented as Anaheim’s. And since one of these teams has to win, we’re picking the Ducks.
If you want to see who Ian’s daughters like for Round 3, you can find out here.
The Department of Hockey Analytics employs advanced statistical methods and innovative approaches to better understand the game of hockey. Its three founders are Ian Cooper, a lawyer, former player agent and Wharton Business School graduate; Dr. Phil Curry, a professor of economics at the University of Waterloo; and IJay Palansky, a partner at the law firm of Armstrong Teasdale, former high-stakes professional poker player, and Harvard Law School graduate. Please visit us online at www.depthockeyanalytics.com
Dr. Mikal Skuterud is a professor of economics at the University of Waterloo.