On Saturday, October 10th, some of hockey’s best statistical thinkers gathered in Rochester, New York to share their current research.
And, as John Matisz put it in an interview with TSN, for many writers from the blogosphere, these conferences are a chance to put a real face to the Twitter handles they’ve engaged in discussion with online, 140 characters at a time. With topics ranging from player evaluation, Corsi talk, the salary cap, injuries and wellness, manual tracking, nationalities, and more, there was something at this conference for every fan. Here, I’ll try to break down a few of the key takeaways.
Analytics is growing and you can’t stop it
The opening panel talk showed that the advanced stats field is gaining real traction in the hockey community. The panel’s speakers were Sam Ventura (recently hired by the Pittsburgh Penguins), Andrew Thomas (recently joined the Minnesota Wild), Matt Pfeffer (employed by the Montreal Canadiens and Hockey Canada), and Timo Seppa (editor-in-chief at Hockey Prospectus).
As of last February, Pfeffer and Seppa were in their current roles. But for Thomas and Ventura, their foray into employment with the league happened recently.
For those in attendance hoping for a shot at NHL employment, the past year of signings (including Ventura, Thomas, and many other bloggers and statisticians) offers real hope that the NHL is waking up to the potential of finding competitive advantage through hiring staff from non-traditional or online ranks.
But the growth isn’t just about people getting jobs.
This conference showed the field of analytics itself is building on to its old principles and fostering the growth of new perspectives.
During the day, listeners were treated to talks focused on some pillars of analytics thought — shift length (presented by Micah Blake McCurdy) and puck possession (David Johnson). McCurdy showed that the majority of player shifts start on-the-fly, while Johnson demonstrated the power of stats relative to team for Corsi For per 60. Johnson also added that the analytics community is still hunting for a good way to evaluate defensemen because “we’re bad at it.”
Along with these talks, presenters offered interesting perspectives on the role of injuries in hockey (Brad Stenger), the impact of the salary cap on GM decision-making (Alex Mandrycky), optimal cap dollar spending (Carolyn Wilke), pass tracking visualizations (Stephen Burtch), thoughts on fusing together qualitative and quantitative data (Stefan Wolejszo), roster “chemistry” (Michael Boutros), and a highly-engaging talk by Nick Mercadante on new stats for evaluating goalies.
Taken together, one thought is clear: the advanced stats field has blossomed, seeking out new layers and uncovering new avenues of study.
Analytics folks are humble…and hilarious
Though the stats crowd carries a reputation for being numbers-focused, a day spent with this group shows off the humour, fandom, and humility prevalent in the field. In fact, Ventura opened the conference with a thought of presenting work humbly that carried through the day:
Matt Pfeffer added the most sobering thought of all:
These humble introspections show a clear development in the hockey analytics groups. New stats aren’t meant to replace or overthrow all of the traditional understandings of hockey. Rather, advanced stats offer refinements and clarity to the way we watch hockey – at best, these refinements are clues. At worst, the insights are just plain wrong.
Though jesting, these qualifiers suggest that the analytics field is now comfortable in its own skin and willing to turn a critical eye on its own understandings and propositions.
Before I wash over you with the sense that the conference was one long moment of zen, know that the day featured the kind of comedy that makes a Saturday spent in a lecture hall feel like a grand idea.
We learned that McCurdy can dance.
And we saw in person that Jen Lute Costella is Melissa McCarthy-levels of funny. She came out firing before her presentation began.
Then dropped quote after quote that left the audience somewhere between amazed and amused.
And my favourite, “GUS NYQUIST DOESN’T NEED YOUR ASSISTS. HE DOESN’T.”
Though her work featured highly interesting player classifications based on goal tracking for more than 7,000 goals (and 5,000 double check views), Jen LC’s enormous data set and interesting insights were sprinkled with the kind of humour that makes heavy stats work entertaining, while still providing the information needed to make the work highly valuable.
Microstats are the way forward and you can help
The last theme was a fairly clear indication throughout the day of where the advanced hockey stats field is headed.
Conference coordinator Ryan Stimson leads the Passing Project, which compiles detailed tracking of every pass for every team in every game over a given sample. Despite the mind-blowing nature of this voluntary enterprise, Stimson has a legion of dedicated team followers who track data and make the passing project a real and valuable source of data.
Jen LC employed a group of approximately 100 volunteers to track the goal data for her presentation. Burtch presented on network analysis, examining the passing connections between a team’s players to identify “nodes” – key players that drive puck movement for a club.
In each case, volunteers are now sifting hockey grains through the finest of filters – tracking every pass and movement across and into every zone. As the presenters noted frequently, more data (and therefore, more help) is needed to compile these kinds of stats. The NHL doesn’t track these event but the analytics crowd is stepping in to fill this void.
Whether or not puck tracking technology eventually will make this data available isn’t yet clear – in the meantime, fans owe much to the work of dedicated team followers who compile highly useful, labour intensive bits of information.
Think you can help? Contact Stimson (@RK_Stimp).
With each passing conference, the hockey analytics movement inches closer to mainstream attention. John Matisz wrote an RITHAC summary for the Toronto Sun on Sunday (here) and the conference was discussed on TSN’s hockey analytics podcast (here). Advanced stats in hockey are here to stay – and they’re growing by the day.
It’s impossible to touch on every presentation but every presentation was valuable. Here’s a list of all of the presentations:
Check back on the conference’s website (here) for the slides used by every presenter. Also check for these folks on Twitter to keep up with their latest research.