A belated review of the SSAC

I had said after the Sloan Sports Analytics Conference that I would write a more detailed summary of the conference, but let things slide until a couple of readers reminded me of what I had promised.  It's been a couple of weeks and my initial feelings are a little fuzzy now, but here are my thoughts on this year's conference.

This is the fifth year of the SSAC, and the second time that I've attended.  Now I didn't attend the first three conferences that were at the MIT campus, but this year's conference is unrecognizable from last year's event.  For one thing, it is now a two-day event (as I suggested last year).  Second, there are more events and exhibits throughout the convention center than before.  Third, the conference is wall-to-wall people…literally!  In the main room where the opening session and the other major panel sessions were held, attendance was standing-room only — and this was a room with seating for almost 1,200 people.

As with previous years, the panel discussions made up the bulk of the conference.  Malcolm Gladwell (author of Outliers) started the conference with the keynote panel on the 10,000 hour rule and its place in the development of elite athletes.  Maybe it was me, but I didn't get a lot out of this session.  I talked to some other people about the session, and most liked it.  So maybe it was me.  At any rate, I walked outside to join the sea of humanity in the hallway.

Developing the Modern Athlete in 10,000 Hours session at the SSAC. (L-R) Justin Tuck, Daryl Morey, Malcolm Gladwell, Mark Verstegen, Jeff Van Gundy.

The hallway was where the action was at this year's SSAC.  Last year there were about five or six vendors representing sports data companies, one or two business analytics companies with a sports subdivision, and ESPN.  This year the vendor booths were expanded to include start-up companies that focused on certain areas of sport analytics.  I've already mentioned the presence of StatDNA at the conference and all of the work that they're doing to promote the development of soccer analytics.  Another soccer-related booth was Chimu Solutions,  a start-up from Prof. Luis Amaral's lab at Northwestern University that developed the Footballer rating system that won a lot of press during the World Cup.  They partnered with a German company that provided them access to live video data, which was exactly what they needed to get more accurate results from their rating.  I spent close to an hour at each booth chatting with the representatives on the products, on the state of soccer analytics, and where we might be able to help each other (within reason, of course).  I had much more substantive discussions with club and company officials in the hallways that I heard in the conference rooms.

That observation brings up another development with the conference and the increased attention.  While high profile officials continue to attend the conference — at least half of the NBA teams sent representatives to the SSAC, even more from Major League Baseball, and significant numbers from the NHL, NFL, and the English Premier League — the panelists who represented those clubs were tight-lipped on their uses of analytics.  There was some more candor in the sessions that didn't have much to do with performance analytics, such as the ones on labor and tax issues in sport, or athlete branding and marketing, but I noticed that fewer decisionmakers were willing to say too much in a public forum with the press — not to mention league counterparts — in the audience.  It's not too surprising.  All of the databases and tools developed by these teams have enormous amounts of value, and in leagues where a 2-3% shift in performance means the difference between a league championship and missing the playoffs, team officials will be very unwilling to say too much, not unlike a hedge fund manager's statements about his strategies at a public forum.  The added attention from the press is another factor.  Mark Cuban, the owner of the Dallas Mavericks, mentioned that the NBA had contacted him before the start of the Referee Analytics session to remind him that he was attending a public forum and could still get fined for comments he made on NBA officiating.  I have no idea if he was serious or not, but it would not have surprised me in the slightest if the NBA were indeed paying attention.

The Referee Analytics session. (L-R) Mike Carey, Mark Cuban, Jon Wertheim, Phil Birnbaum, and Bill Simmons.

Next to the exhibitor displays and the hallway interactions with attendees, the most useful part of the conference were the poster and research paper presentations.  Unfortunately I did not attend any of the paper presentations, so I can't comment on those, but I did attend the poster presentations.  The paper and poster presentations were dominated by basketball research, from in-game strategies to player performance in contract seasons to refinements of the adjusted plus/minus rating.  I liked the poster on adjusted plus/minus extension to hockey as well as the research on player evaluation using a finite state machine model.  I am impressed with the depth and breadth of basketball research, and I'm more than a little envious that basketball enjoys a big database of play-by-play data that is suitable for analysis.  Even with the overtures that Opta and other companies have made, soccer is well behind baseball and basketball in this regard.  You can read the SSAC research publications here.

I didn't hear anything about a best poster award, but there was a best paper award given to an economics student at Boston University on player performance in the NBA just before a contract expires and immediately after signing a new long-term deal.  The winner received a $7500 award from Titleist, which goes to show the increased sponsorship of the conference and the amount of money that is flowing into analytics work. 

There was also an Evolution of Sport (EOS) session, which was modeled after the TED talks in which someone has what they think is a cutting-edge idea and is invited to give a brief talk on it in a manner that is supposed to be engaging and insightful.  I attended one of the talks, which a friend of mine (the founder of Sports Data Hub) gave and was actually quite good.  It was short — less than five minutes! — which I liked most of all.  People's mileage will vary on the EOS talks, but reading some of the abstracts I can understand the general public's unease with sports analytics in particular and "smart" people in general. 

Making over-generalizations is not a good idea, but Ivy Leaguers tend to hold themselves and the ideas that they espouse in particularly high regard, and this conference is full of both.  When I hear presenters talk about the future of analytics as if we're going to have robotic replacements for managers, or even robotic (or cyborg-like) replacements for players, my eyes start to roll.  I'm pretty sure that a layperson's eyes will start rolling as well, and the credibility of the analytics community takes a hit.  I also find it difficult to not laugh out loud when some reporter writes that the future of sport analytics will be some number cruncher sitting beside the manager or coach during the game (before or after is a different story, and is already happening).  I think that sports analytics are here to stay, but it would be great if we in the analytics community don't get so wrapped up in how brilliant we and the tools we develop are.  There is still a fair amount of common sense residing in the minds of the old-school scouts and coaches.  A lot of the thinking might be outdated and needs to be changed in light of what we better understand, but it doesn't mean that it should be discarded altogether like it's Year Zero.

I did notice that the conference is very Anglo-Saxon, with some Germanic elements thrown in.  It is understandable that it be so, since most of the sports analytics work being done is in either the USA or the UK (there is soccer research going on in Continental Europe, but I don't know if the clubs there make use of it).  Perhaps they do and I'll find out more about their work when I attend the Leaders in Performance meeting later this year.  But I would love to see a greater participation from researchers in countries outside North America and the British Isles. 

So now if you've read this far, you now know what I thought of this year's conference.  I will be there for sure next year, but I think I'll spend more of my time in the hallways next time.  I'll leave you with the now-famous Mark Cuban shirt: