Randomness Revisited: A Look Back at the Penguins’ 15-Game Winning Streak

There were a lot of reasons why the Pittsburgh Penguins entered last year’s playoffs as the top seed in the Eastern Conference, but the most obvious has to be the distance they put between themselves and everyone else during the month of March. They didn’t lose a game, either in regulation or in overtime, for the entirety of that month. While the Penguins are obviously a good team, and good teams do go on streaks, things like a 15-game winning streak just illustrate how hard it is to predict the outcomes of hockey games. When you get into small samples, lots of random, unexpected things happen. Consider the following:

  • The streak began with a 7-6 overtime win against the Canadiens. Since 2005, only 2.6% of teams allowing six goals in a game have gone on to win that game. But sometimes you just score on 19% of your shots and win anyway.
  • The third game in the streak was a 5-4 win in Philadelphia. The first period of this game was arguably Pittsburgh’s worst period all season, as Marc-Andre Fleury gave up four ugly goals and the team needed to come back from 4-1 down to win. As with the Montreal game, it’s nice to shoot over 22% as a team sometimes.
  • The Penguins played five of these games without Kris Letang, and 10 of them without Evgeni Malkin.
  • They beat the Boston Bruins (who would sweep them out of the playoffs) twice during this streak, including one game in which they were held to just 18 shots. In the other game, they dominated the Bruins territorially, outshooting them 34-16, despite Boston being one of the NHL’s best possession clubs last season.
  • Nine of the wins during the streak were by 1 goal. Four required a third-period comeback.
  • Pittsburgh’s goal differential during the streak (+26) was greater than their shot differential (+24).

Photo credit: Flickr user Doug Butchy. Use of this image does not imply endorsement.

Speaking of shot differential, the Penguins did not accomplish this streak by dominating puck possession. As you can see from their 2013 chart over at Extra Skater, Pittsburgh’s cumulative Fenwick Close % actually declined over the streak. (If I’m speculating, I’m pinning that squarely on Malkin’s absence from the lineup.) Using Extra Skater to look at things by game, their possession statistics were all over the place. (Some people will tell you that Fenwick Close is a consistent, repeatable measure of possession ability; I view that as more of a hypothesis than a fact.) In their March 4 and March 7 wins over Tampa Bay and Philadelphia, the Penguins had a combined Fenwick Close of just 38.2%, and that’s against two lousy possession clubs. In the game immediately following that, they played the Maple Leafs, and had over 77% of Fenwick Close events in a shootout victory. They had another strong possession game against a good team in the Rangers (Fenwick Close 62.2%), but were lucky to escape with a 1-0 win in their second game with the Canadiens (Fenwick Close 32.6%). Generally speaking, the picture these numbers paint is hardly one of a team that was winning by tilting the ice towards the opposing goal; over the course of the winning streak, the Penguins’ Fenwick Close was just 49.9%

So how did Pittsburgh manage to pull this off? Well, as is so often the case, the explanation comes back to the PDO statistic: over the streak, the Penguins recorded a staggering 1057 PDO. Interestingly, their team shooting percentage of 10.6%, though excellent, isn’t all that remarkable for a team on a winning streak. Where the Penguins were killing it, oddly enough, was in goal: the team’s even-strength save percentage was 0.951 during these 15 games. Fleury and Tomas Vokoun pitched four shutouts during the streak, and gave up a total of six goals in the final nine games of it. Clearly, neither Vokoun’s 0.946 Sv% nor (especially) Fleury’s 0.954 were sustainable in the long run. Which is something that should give every fan pause: no team is quite as good as they look when they’re on a winning streak.

So, how unlikely is a 15-game winning streak, exactly? Well, if all other things are assumed equal, the probability of winning 15 games in a row is about 1 in 33,000. In an 82-game season, each team has 68 15-game sequences in its season, which gives us roughly 1,000 sequences of 15 games per year in a 30-team league. This implies that a 15-game win streak should occur about once every 30 seasons. Of course, the NHL hasn’t always featured so many teams or such a long regular season; since 1967, the number of teams has grown from six to 30, with the number of games per season ranging from 48 (in lockout-shortened years) to the current 82. As such, win streaks of this length have not happened predictably at 30-year intervals: the Penguins’ streak last year came almost exactly 20 years after their 17-game streak in 1992-1993, but the only other 15-game winning streak in NHL history belongs to the Islanders in the winter of 1982. Which also means that it took 64 years for the first win streak of this length to occur. What’s more, as with any calculation involving, you know, exponents, winning streaks become exponentially less likely the longer they run. The Capitals’ 14-game streak in 2010 was certainly impressive, but the probability of that happening is about 1 in 16,500, or twice as often as a 15-game streak. The Sharks started this past season winning seven straight, and cruised into the final stretch with another seven-game streak in early April. This was certainly cool, but it’s 128 times more likely than what the Penguins did. So, call it lucky, call it random, call it whatever “peaking too soon” narrative you like: things like the Penguins’ month of March are why the NHL is never boring.

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 Original Analysis and tagged , , , . Bookmark the permalink.

One Response to Randomness Revisited: A Look Back at the Penguins’ 15-Game Winning Streak

  1. Pingback: What We Can Predict, and What We Can’t | 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