Todays SlapShot

15 June 2011: The Boston Bruins celebrate winning the Stanley Cup at the end of Game Seven of the Stanley Cup Finals between the Vancouver Canucks and the Boston Bruins held at Rogers Arena in Vancouver, British Columbia, Canada. Final Score Bruins 4 Canucks 0 - Bruins win the Stanley Cup ****For Editorial Use Only****
Boston Bruins

How the Bruins won the Stanley Cup without analytics

Bob Frid/Icon Sportswire

Vancouver Canucks general manager Jim Benning comes from not-so-humble managerial beginnings as the Assistant General Manager for the Boston Bruins.

While in Boston, the former defenseman saw his name engraved on the Stanley Cup for the first time – and he and the team did so, as he reminded us all this week, without analytics.

“I’ll be perfectly honest with you,” he announced, following the deal that brought defenseman Erik Gudbranson to Vancouver, “We won a Stanley Cup in Boston and we didn’t use analytics.”

This isn’t the first time we’ve seen stats dismissed by those who think they don’t matter. We’ve heard it in the past in basketball, as well. After all, advanced stats weren’t really around when Michael Jordan and the Chicago Bulls ran roughshod over the NBA almost 20 years ago.

If they could do it, why would you assert that you need to use stats to create a dynasty in the modern era in sports, as well?

To an extent, this is accurate.

Advanced stats do not ‘make’ players good in the same way that skills development can. Yes, you can tell a player that his shot suppression metrics are a travesty, and he’ll be able to work to fix that – but Corsi does not describe the ‘how’ of a player’s talent. It simply quantifies it as best as numbers can quantify something that’s part skill, part luck.

Benning makes an egregious mistake, though, when he argues that poor analytics deals are good deals because he’s seen champions built without numbers.

The 2010-11 Boston Bruins may not have been built by analysts, but that doesn’t mean that they weren’t an statistics-friendly team.

19 June 2013: Chicago Blackhawks goalie Corey Crawford #50 tries to get a line of sight around Boston Bruins right wing Nathan Horton #18 during game 4 of the Stanley Cup Final between the Chicago Blackhawks and the Boston Buins at TD Garden Boston, MA

19 June 2013: Chicago Blackhawks goalie Corey Crawford #50 tries to get a line of sight around Boston Bruins right wing Nathan Horton #18 during Game 4 of the Stanley Cup Final between the Chicago Blackhawks and the Boston Bruins. Photographer: Michael Tureski/Icon Sportswire

The best offensive possession player on the Cup-winning Boston Bruins, in all situations, was power forward Nathan Horton. He put up a whopping 59.96 Corsi For percentage, making him a monster possession driver.

His shot metrics, measuring the number of shots his team took per sixty minutes with him on the ice versus the number of shots allowed, were even more impressive. In all situations, Horton saw his team take 37.8 shots on goal per sixty minutes, while they only allowed 27.6 shots.

That particular Boston roster may not have been built using analytics, but it was an absolute possession dream.

Patrice Bergeron, known as one of the NHL’s best two-way centers and a perennial Selke Candidate (widely considered an elite possession player by the advanced stats crowd), was only the seventh-best player on that Bruins roster from an analytics standpoint.

The seventh best. Let that sink in.

It’s entirely possible that teams can see analytics-friendly rosters fall into their laps as a happy accident. You don’t need to know a player’s Corsi or Fenwick before they appear in a team’s lineup for them to produce top results; again, analytics just measure the ‘what’, not the ‘how’.

Teams have yet to figure out how to assemble a ‘Corsi Army’, so to speak, without bringing in analytics-friendly players to begin with.

The problem, though, is precisely that – you cannot create a good possession team out of thin air. You can only measure how good the team’s numbers are.

That’s why teams that have embraced the analytics movement have done so. They aren’t writing a formula on a chalkboard that will come to life, seep into the veins of their defensive corps, and win them a championship – they’re bringing players on board with that the numbers already favor.

They’re bringing in the Nathan Hortons and the Patrice Bergerons, the Marc Savards and the David Krejcis, by calculating ahead of time which players have yielded results that best align with the characteristics of teams that have won in the past.

That’s part of the problem with those that still reject analytics.

Jim Benning, it seems, doesn’t fully realize what it is that analytics can do. They can’t turn mediocre players into game-changers; it’s coaching, skills development, and systems utilization that does this.

Using analytics doesn’t mean that your team will automatically have good numbers. For example, take a look at the Arizona Coyotes. This past year, they were a poor possession team.

Hiring an analytics guru to start working with the team’s numbers didn’t teach their young corps how to execute clean zone entries, suppress shots, and execute effective passing rushes. All it did was enable them to take a look at why they weren’t winning (they were an average-luck team with below-average scoring opportunities and possession numbers), and identify which players were causing those problems.

They did that though, weeding out what wasn’t working, and making the jump from a bottom-out season to a not so painful one.

Those numbers do help better predict when a certain deal will quickly sour for a certain team – and help a team determine when they’re playing well due to luck or when they’re sustainably good.

March 03 2016: Colorado Avalanche Head Coach, Patrick Roy during a regular season NHL game between the Colorado Avalanche and the visiting Florida Panthers at the Pepsi Center in Denver, CO. (Photo by Russell Lansford/Icon Sportswire)

March 03 2016: Colorado Avalanche Head Coach, Patrick Roy during a regular season NHL game between the Colorado Avalanche and the visiting Florida Panthers. (Photo by Russell Lansford/Icon Sportswire)

The Colorado Avalanche are a perfect example.

A notoriously poor possession team under head coach Patrick Roy, the team outperformed expectations during the 2013-14 season. They made the playoffs, despite analytics pundits everywhere suggesting that they weren’t actually good, just very lucky.

Sure enough, they’ve been crashing back to earth ever since, falling out of the postseason in both 2014-15 and 2015-16 to the frustrated confusion of their coach and management staff.

Then there’s the Canucks – Benning’s own cavalry of non-believers, who are currently trying to prove that the eye test can make up for what the numbers are suggesting isn’t there.

After all, the Bruins won a Stanley Cup without numbers, so who needs the numbers for Gudbranson, when there’s a gut feeling that he’s what it takes to win a championship?

That logic seems to suggest that if you don’t actively use the numbers, they don’t actually exist.

The Bruins may not have used advanced analytics to build their 2010-11 roster, but that doesn’t mean that the team had no measurable possession impact.

Failing to look up a player’s shot metrics doesn’t mean that they have no shot metrics – it simply means that you, as the management staff, weren’t aware of what those metrics were when you deployed the player.

Sometimes, this is a fairly moot point. You don’t need to look at a possession chart to know that Bergeron has a tangible impact on his team’s ability to win games. You don’t need to look at fancy stats to know that Patrick Kane, Jonathan Toews, Sidney Crosby, or the Sedin twins are some of the best players in the game.

In many instances, what your eyes are telling you is exactly what the stats will reaffirm.

In murkier areas, though, the stats can tell you what your eyes are missing.

Coyotes head coach Dave Tippett, one of the more analytics-friendly coaches in the NHL, pointed out during the 2015-16 season that hits were a ‘noisy energy’ stat.

While fans of heavy-hitting hockey assume that hits make a team harder to play against – not altogether wrong from a physical standpoint – they miss that teams don’t deliver big hits when they already have the puck. Finishing a game with a ton of hits can often mean nothing more than that a team spent the majority of the night without the puck.

Heavy hitters and players who stay in the defensive zone are sometimes argued as guys good at breaking up plays. They’re good at shutting down the opposition, the anti-stats pundits argue, and there is no stat for that.

To Jim Benning, there is no way to measure what he believes to be Gudbranson’s biggest asset, and therefore, his value is impossible to determine.

A player’s defensive value is actually not impossible to calculate, though — it just isn’t used in Vancouver.

Shot suppression can be actively measured, and Gudbranson’s ability to limit shots against his goaltender has been on par recently with oft-criticized defensemen like Luca Sbisa and Jared Cowen.

Jim Benning is right. He doesn’t need analytics to pick up good players.

Sometimes, he will be able to acquire those just by looking at their on-ice performances. And while he didn’t use the numbers to decide the players are good, he will have acquired a good player regardless. The metrics don’t make the player good, they simply show that he is.

When the numbers suggest a player has noticeable holes in his game, though, that’s not something that can be ignored by saying that other winning teams have won without looking at the numbers — because the numbers were always there.

How the Bruins won the Stanley Cup without analytics

Comments
To Top