Last week I attended my first MIT Sloan Sports Analytics Conference as a student attendee. Sloan is a lot different than most conferences that would be relevant to someone in Operations Management. For one, it only accepts 8 research papers for presentation. For two, most of its presentation slots are panels where the participants come from industry. Specifically, sports teams or companies.
Overall, I had a good time. Most of the research paper talks I attended were good, and maybe half of the panels I attended said something interesting or funny. It’s a common refrain from attendees that panelists are trying to be intentionally vague or misleading so that they don’t give away any competitive advantage. While I understand this, it makes for some boring panels.
I got to hang out with Charles Glover, a former co-worker at Booz Allen that was at the conference showing off Booz’s data science capabilities. Booz Allen co-sponsored the conference enough to have a permanent “data visualization” zone in one of the conference rooms. Apparently Booz Allen is helping run Major League Baseball’s replay headquarters since last year’s introduction of coaching challenges. Here is an overview of how Booz Allen is using data science in sports.
I queried Matthew Berry, resident fantasy expert, for tips for Maria’s new class. During the next January term at DePauw, she will be teaching “How to Use Data Science to Win at Fantasy Football”. That should be fun. Berry had suggestions for looking at draft value by position and number of draftees still in lineups in Week N.
I wrote a few notes from some of the interesting panels and presentations, which I’ll share here:
Michael Lewis (author of Moneyball and The Blind Side) was on a panel with Daryl Morey (Rockets GM), Jeff Van Gundy, and Shane Battier. Lewis had a few good lines, including “You can’t be too stupid to play baseball”. He said that he wanted to interview Battier for this article because Battier was a lab rat who could understand the experiments his coaches and GM were putting him through. Battier, for his part, had a good line about Heat coach (misspelled) Erik Spoelstra: “Spoelstra told me, ‘Don’t dribble, don’t post-up, don’t offensive rebound. Just catch and shoot or catch and pass'”. Battier’s teammates LeBron, Wade, and Bosh got a little more freedom, I imagine.
Mark McClusky, from WIRED magazine, shared some interesting thoughts on the evolution of performance in sports. “The only competitive advantage is to learn faster.” Once you implement a strategy or an insight, others will copy it, so you need to keep learning to stay good. He also shared some research on sleep and suggested that everyone needs to sleep more to be at peak performance. Some suggested books for reading include Better, Faster, Stronger by McClusky, The Sports Gene by Epstein, and Better by Gawande.
There was a new fantasy platform called Reality Sports Online that will begin a big advertising campaign to get players this year. It is a more complicated/intense version of fantasy involving mullti-year contracts, negotiations, and rookie drafts. It’s meant to mimic the general manager experience more than current fantasy leagues.
Dan Rosenheck, writer at the Economist sports blog, talked about how Spring Training statistics had some predictive value, contrary to popular belief. A brief writeup of his presentation is here.
Brian Burke kept his cool in a silly football analytics panel, which was impressive. He said that an NFL game boils down to 11 minutes of gameplay, which means a defensive coordinator could watch all of his squads plays and grade his players 5 or 6 times in half an hour. As such, it could be years before player tracking data beats insights from tape watching.
Boston was cold and snowy. I don’t suggest visiting in February for tourism.
Other Recaps of SSAC 2015:
What it’s like to be a woman at SSAC
Soccer Analytics Panel at 2015 SSAC: Not a waste of time
2015 Sloan Recap: Where are the Analytics?
SSAC 2015 Takeaway – Accepting Yes For An Answer
The Value of a Good Analytics Program by Brian Burke
. This saves you from having to find and click on all the suggested symbols in Word to get an equation you want.