2013-14 NHL Playoffs: First Round Predictions

At long last, the NHL postseason is back, and Puck Prediction has you covered for the coming weeks. I’ve described my method for predicting the series winners before, but the gist is this: I’ve combined a measure of each team’s overall quality, a modified version of GF% that I call mGF%, with the effects of home-ice advantage to construct an analytic model that offers conditional probabilities of each team winning their series. Now that I’ve got 82 games of data on all 16 teams, let’s get started by previewing and picking the first-round match-ups.

Atlantic Division

Boston Bruins (mGF% 55.7%) vs. Detroit Red Wings (mGF% 51.1%)

playoff_BOSDET_1

Both of these teams prevent shots very effectively, and insofar as I don’t think the Bruins’ scorching-hot shooting down the stretch is a sustainable phenomenon, my expectation is that this series will be a low-event slugfest, in which neither team decisively controls possession. The real game-breakers, for me, are home-ice advantage and Tuukka Rask. Jimmy Howard has struggled with injury at times this season, but has quietly had a strong campaign (0.925 Sv% at even strength). Yet Rask’s 0.942 is the NHL’s best among goalies who played at least 41 games this season. As 1-8 match-ups go, this is far from a cakewalk for Boston, but I expect that they’ll come out ahead.

Prediction: Boston in five games.

Extra Skater stats dashboards: Boston, Detroit

Tampa Bay Lightning (mGF% 53.2%) vs. Montreal Canadiens (mGF% 49.6%)

playoff_TBMTL_1

Much depends on the status of Tampa starting goalie Ben Bishop, who had a magnificent season prior to an injury sustained in the season’s final week. If he’s unable to play (or play effectively), the fallback plan is Anders Lindback, who struggled to a 0.914 Sv% in 23 games this season. In contrast, Carey Price has been magnificent in the Montreal goal this season. Which is a good thing: only the Oilers, Sabres, and Maple Leafs allowed more unblocked shot attempts against per 60 5-on-5 minutes than the Canadiens’ 60.4. The Lightning, however, were a top-10 shot-prevention team this season. As such, I’d expect Tampa Bay to control puck possession consistently in the series. Both teams are loaded with offensive skill up front. As has been the case all season, Montreal’s hopes will rest on Price’s ability to weather the storm in front of him.

Prediction: Tampa Bay in five games.

Extra Skater stats dashboards: Tampa Bay, Montreal

Metropolitan Division

Pittsburgh Penguins (mGF% 52.1%) vs. Columbus Blue Jackets (mGF% 50.5%)

playoff_PITCBJ_1

I suppose this series is the sort of thing the NHL had in mind with this season’s division realignment: despite their geographic proximity, the Penguins and Blue Jackets have never played a meaningful game against one another. This series, however, should help to create more of a rivalry. While Pittsburgh has become a regular playoff performer in the Sidney Crosby/Evgeni Malkin era, Columbus is making just its second postseason appearance ever; they’ve yet to win a playoff game as a franchise, having been swept by Detroit in their only appearance in 2009. Nevertheless, the match-up here is not as one-sided as you might think. Both Columbus and Pittsburgh are low-event teams with slightly positive possession numbers, so it’s far from obvious that the Penguins will drive the play in this series. In goal, reigning Vezina winner Sergei Bobrovsky has been very strong for the Jackets (0.933 Sv% at even strength). In contrast, Penguins fans will be on pins and needles as Marc-Andre Fleury attempts to exorcise the demons of his last two first-round series. Fleury had an unspectacular 0.919 Sv% this season, and Pittsburgh may not have the luxury of turning to Tomas Vokoun if he falters again. Still, one thing worth noting about the Penguins is that they’re entering the playoffs largely healthy. If Pittsburgh’s defense corps and Crosby and Malkin up front are ready to go, the challenge may be too much for Columbus.

Prediction: Pittsburgh in seven games.

Extra Skater stats dashboards: Pittsburgh, Columbus

New York Rangers (mGF% 54.2%) vs. Philadelphia Flyers (mGF% 48.6%)

playoff_NYPHI_1

I know the Flyers defied my expectations this season, overcoming both their poor defending and 61 starts of Steve Mason to make the playoffs. But this is a very difficult match-up for them. New York is one of the best puck possession teams in the NHL, and Henrik Lundqvist’s shaky start to the season seems like a distant memory. Even if Mason can keep his solid season going, the Rangers still have a clear edge in goal, and should outshoot the Flyers by a healthy margin in the series.

Prediction: New York in five games.

Extra Skater stats dashboards: NY Rangers, Philadelphia

Central Division

Colorado Avalanche (mGF% 48.6%) vs. Minnesota Wild (mGF% 50.0%)

playoff_COMN_1

I’m glad to have a model to make a call on this series, because it’s very difficult to see where this match-up is going. On one hand, my expectation is that Minnesota will control puck possession against Colorado: the Avs are one of the NHL’s most porous defensive teams, while Minnesota is one of the stingiest. While few have forgotten his disastrous tenure in Philadelphia, I’m not especially concerned about Ilya Bryzgalov’s ability to succeed behind such a stout defense. On the other hand, the Wild’s offense has been utterly anemic this season: only five NHL teams scored fewer goals this season, and only the hapless Sabres recorded fewer shots on net than Minnesota. As such, it’s debatable whether they can score enough to outlast the Avalanche in this series, especially with likely Vezina nominee Semyon Varlamov in the crease for Colorado.

Prediction: Minnesota in six games.

Extra Skater stats dashboards: Colorado, Minnesota

St. Louis Blues (mGF% 52.4%) vs. Chicago Blackhawks (mGF% 54.2%)

playoff_STLCHI_1

The Blues may come to regret the six-game skid at the end of the season that pushed them out of first place in the Central: instead of facing a Wild squad that they’d likely have beaten, they get a brutal match-up with the defending champion Blackhawks. Chicago and St. Louis probably have the two deepest bluelines in the NHL, and both are exceptional shot-prevention squads. As such, this is likely to be a very low-event series. After a shaky start, Corey Crawford has rebounded to have a solid season (0.927 Sv% at even strength). Ryan Miller’s performance this season has been a bit harder to read: after playing well in front of an abysmal defense in Buffalo, Miller has been fairly average since his trade to St. Louis. This series is a bit of a coin flip, but I’d have to guess that Chicago’s superior possession game gets them through.

Prediction: Chicago in six games.

Extra Skater stats dashboards: St. Louis, Chicago

Pacific Division

Anaheim Ducks (mGF% 53.9%) vs. Dallas Stars (mGF% 51.7%)

playoff_ANADAL_1

As a match-up between two high-event teams, each featuring a dynamic scoring duo (Seguin/Benn in Dallas, Getzlaf/Perry in Anaheim), this series should make for good television. Neither squad is particularly good defensively, so I don’t expect the Stars to have much (if any) possession advantage. Kari Lehtonen and Jonas Hiller are both capable goaltenders that could steal this series if they can find a run of hot play. It’s impossible to say whether Anaheim’s 9.8% shooting will show up in these games, though they’ve kept up a high Sh% for so long I’d be reluctant to bet against it.

Prediction: Anaheim in five games.

Extra Skater stats dashboards: Anaheim, Dallas

San Jose Sharks (mGF% 53.7%) vs. Los Angeles Kings (mGF% 52.9%)

playoff_SJLA_1

Sharks fans are feeling pretty pessimistic about this match-up because, frankly, pessimism is how we approach postseason hockey. If you recall, the Kings and Sharks played a tense seven-game series a year ago that ended with a 2-1 victory for the home team at Staples. Since then, LA has put together another season of dominating possession hockey, leading the NHL with a 56.7% Fenwick Close, and getting strong goaltending from Jonathan Quick (0.928 Sv% at 5-on-5). Unsurprisingly, the Kings were the toughest team in the league to score on this season. As such, the arguments for optimism in San Jose look something like this:

    • If any team has an offense dynamic enough to break down the Kings, it’s the Sharks, especially with a healthy Tomas Hertl on the ice.
    • There’s a reason the Kings fell to third in the Pacific after a long skid in December and January: no team in the playoffs this year scored fewer goals than LA, and only the Buffalo Sabres had a lower even-strength Sh% than the Kings’ 6.6%. If they can’t find the back of the net against Antti Niemi, the Kings’ presumed advantage in possession isn’t going to matter.
    • Quick is a career 0.915 goalie. At some point, the Statistical Regression Fairy is bound to point in his direction in a playoff series.

Prediction: San Jose in seven games.

Extra Skater stats dashboards: San Jose, Los Angeles

About Nick Emptage

Nicholas Emptage is the blogger behind puckprediction.com. He is an economist by trade and a Sharks fan by choice.
This entry was posted in Predictions and tagged , , , . Bookmark the permalink.

8 Responses to 2013-14 NHL Playoffs: First Round Predictions

  1. I used season series GF% of the higher seed and momentum (pts earned in last 10 games of the higher seed minus pts earned in L10 of underdog). Data was used from the past 16 seasons. I ran a multiple regression on it. The intercept term can be interpreted as home advantage. Higher seed percentages to win the series according to my model: BOS 52.83% Pit 68.8% TB 63.86% NYR 69.58% ANA 63.08% COL 71.73% STL 45.38% SJ 51.56%
    Main difference in our models is with Colorado. Boston is expected to be closer in my model and Pittsburgh expected to be closer in your model. The other 5 series are pretty close

    • Nick Emptage says:

      Cool. I would worry about focusing on season series because the data sample is so small, but this is really creative. And Josh Weissbock uses a version of your “momentum” factor, but he focuses on Fenwick events rather than standings point.

  2. omgitsdomi says:

    How accurate has this model been in the past?
    I figured one out that was 75% efficient over the last 5 years (small sample) combining fenwick close, goals for % adjusted for PDO, and special teams data. Only difference I had was Colorado instead.
    Very interested to see how the two compare

  3. Drew says:

    I have Colorado winning the series, but I could see how you’d go MN… NHL is tough to predict. I really hope Columbus pulls out a first round win vs. Pen15′s.

    • Nick Emptage says:

      COL-MIN is basically a coin flip, but I had to pick someone. I think Columbus will give Pittsburgh all they can handle; if Malkin can’t go and Fleury struggles the Jackets could definitely win that series.

  4. Pingback: Understanding Home-Ice Advantage in the NHL Playoffs | Puck Prediction

  5. Pingback: 2013-14 NHL Playoffs: Second Round Predictions | Puck Prediction

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s