This year was my third time attending the Sloan Sports Analytics Conference, and it’s amazing to see how it has changed in such a short time period. Baseball has become increasingly advanced, and basketball has arrived at the forefront of the sports analytics movement. Here are some of my notes from the conference on the things I found most interesting:
Pushback Against Analytics
Even if it’s the statistically right decision, in-game tactics can often create a backlash of externalities when they fail. One of the most interesting panelists at the conference was Kevin Kelley, the high school coach who goes for nearly every fourth down and onside kick. In the NFL, coaches who have made crucial 4th down decisions get crucified if it fails, even if the odds were in their favor. The panel on In-Game Innovations had the perspective of coaches George Karl and Kelley, statisticians Bill James and Nate Silver, and Daryl Morey from the front office.
As the general public, players, and front offices get smarter and smarter, these effects will decline. However, they still affect decisions in 2014 – despite the facts found about 4th down efficiency, NFL teams did not go for it any more than previous years. Nate Silver called a gimmick “something so far ahead of its time that there’s no way to know whether it’s effective or not.” As time goes on, we’ll continue to develop better predictors of anything that happens on the field.
Technology & New Data
We’re all familiar with sportVU at this point, and the ability to quantify things like field goal percent against a defender at the rim explains much of what was not quantifiable before. My favorite research paper was Kirk Goldsberry and co.’s on expected possession value (Grantland article on it here and full paper here). The scope of possibilities with the ability to track the likelihood of any single event on the court is nearly limitless. The challenge now is to figure out what is relevant and what is not.
This research and data isn’t limited to basketball either. MLB Advanced Media presented a slew of new data that will be a part of broadcasts and statistics from now on, and I’m not the biggest soccer fan, but XY tracking for soccer will lead to a whole new dimension of analysis for the game.
Progress in Analytics
Last year I listed five new areas of analytics that needed more research, and much of the conference this year was geared towards some of what I highlighted. Last year I listed Injury Analytics, Scouting Analytics, Coaching Analytics, Team Chemistry Analytics, and Spatial Analytics. SportVU made spatial analytics a focus of the conference, while injuries, coaching, and team chemistry remain very hard to quantify. The focus of the conference is always a combination of advanced statistical work and looking for the easiest ways to translate this research into usable and logical statistics for players, teams, and fans alike.
When in Doubt, Use Blind Resumes
One of my favorite panels was one revolving around the new college football playoff system. It had the audience participating and voting on the resumes of teams in past years who would have been in contention for the fourth and final playoff spot. Much of this same logic can be translated to fantasy drafts and trades, where name values can skew a player’s real value. For the upcoming baseball season, going in to your research open to new values is the easiest path to success, and the easiest way to find this year’s hidden gems. Taking away the inherent biases in normal analysis is an easy way to make better decisions.
- Not only is Stan Van Gundy hilarious, but he made some really good points on the Basketball Analytics panel. He talked about a main flaw in effective field goal percentage, and why 33% on 3′s isn’t the same as 50% on 2′s. The value added from the 50% of 2′s is increased because the made basket allows less transition opportunities for your opponent. Van Gundy also talked about why injury limits are frankly useless, and injury analytics are still far behind many aspects of the game. He cited Strasburg as an example, where 160 innings was an arbitrary endpoint and the Nationals really needed him in the playoffs.
- While they didn’t add much in terms of analytics, the panel on “Building a Dynasty” with Phil Jackson and Jonathan Kraft and the 1 on 1 interview between Malcolm Gladwell and Adam Silver were both incredible opportunities to hear some interesting stories and look towards the future of the NBA.
- I missed the “Athlete Analytics” panel but a repetitive theme throughout the conference was that successful analytics need to be easily communicated. Niners’ president Paraag Marathe said, “Data is great, analytics are great, but communication is the most important.” That doesn’t mean to damper more advanced analytics – it just means your findings need to be understandable and concise.
Matt Cott is a co-founder of RotoAnalysis and writer for CBS Philly and The New York Times. Follow him on Twitter @KidCott21!