2014 MIT Sloan Sports Analytics Conference – Day 1
Ever since I started following and studying the NBA to the degree at which you must to write about it knowledgeably, I have wanted to attend the MIT Sloan Sports Analytics conference. Throughout the eight years of the conference’s existence, though, tickets have become increasingly sought after (according to the organizers, there was a waiting list of over 1,000 people this year), not to mention more and more expensive (close to $600 for a regular admission ticket). Fortunately, my enrollment in Texas’ MBA program allowed me to not only get a ticket, but get it at the student rate (about a third of the full price). The result was two days in what I like to refer to as “sports nerd heaven.” Day 1 is detailed below, and I’ll go through Day 2 in a later post. If you have any other questions about the conference, feel free to tweet @MasonGinsberg and ask whatever you’d like.
The conference began at 8 AM with a welcoming from the main organizers from MIT, as well as the creators themselves. After the introduction, I was off to my first panel choice of the day which focused on in-game innovations for various sports. This panel ended up being one of my favorites of the weekend, as it consisted of Rockets GM Daryl Morey and former NBA coach George Karl, along with Nate Silver, Bill James, and Kevin Kelley (with TrueHoop’s own Kevin Arnovitz moderating). They discussed multiple interesting topics and proposed some unique ideas. For example, George Karl suggested having a single-elimination NBA tournament during All-Star Weekend, as the mentality from both a player and coaching perspective would be vastly different than the standard best-of-seven playoff format. This idea prompted discussion about parity in the major sports and how the NBA has the least of any of them. Morey attributed this fact to the best-of-seven series, as it allows for less of a chance for the weaker team to upset the better team. It was interesting to hear great minds like Morey and Silver question whether or not the best team almost always winning in the NBA is a good thing for the league. Other notes from this panel:
- George Karl seems to truly believe that the difference in talent between the worst and best team in the NBA is not terribly significant, but that some teams know how to win in the 4th quarter better than others.
- Both Morey and Karl agreed on the notion that slow-playing offenses are inefficient, and that “chaos” leads to points. Quick decisions made before the defense can get set are typically more effective than even some of the best designed set plays.
- Karl wishes that there was a way to quantify the concept of “teamness”, which he said allows for a special chemistry and energy that makes teams better.
- The often debated topic of whether or not to foul when up by 3 at the end of games was discussed, and everyone seemed to agree that it’s the smart play. However, it is something that needs to be practiced to execute it correctly (so not to foul a shooter), which teams don’t frequently do. Additionally, some coaches and players may not want to do it because it gives off the vibe that they don’t trust their own defense.
- Lastly, Morey pushed the idea that the NBA has to “get rid of the marginal incentive to lose.” He mentioned the idea of a flat lottery system (even odds for all non-playoff teams), but this topic wasn’t discussed in greater detail due to time running out.
Next, I moved on to a panel about starting sports businesses. Some of my key takeaways:
- Difficulties dealing with the power of the leagues and league rights deals is one of the biggest barriers for sports business start-ups.
- One of the biggest concerns for sports businesses is counterfeiting and commoditization of products.
- It has never been quicker or cheaper to start a sports business than it is today. Scaling a business is when costs start to pile up.
After lunch, I went to a session about how SAP is improving sports analytics. As far as the NBA is concerned, the system has done wonders in this area by publicizing data, as can be seen at NBA.com/stats. The ability for any NBA fan to see shot charts, heat maps, and game footage (along with every box score in league HISTORY) with just a few clicks is truly extraordinary.
I started my afternoon off by listening to a presentation about a research paper which ended up being named the best one at the conference – “The Three Dimensions of Rebounding.” The analysis done by author Rajiv Maheswaran was really quite incredible, as he broke rebounding down into three phases – Positioning, Hustle, and Conversion – and could quantify player ability in each of these areas individually or in total. Notable findings according to his analysis:
- Among non-big men, Tyreke Evans ranks 5th in the NBA in the ability to establish good initial offensive rebounding positioning.
- In the second dimension of hustle, Jason Smith ranks 6th in the NBA on the offensive end and Al-Farouq Aminu ranks 3rd on the defensive side. Hustle is most simply defined as a player’s ability to turn bad pre-shot positioning into good post-shot positioning.
- The final dimension, conversion, indicates a player’s ability to turn positioning into actual rebounds.
- This analysis values Aminu’s rebounding very highly, and fairly so. Given this fact, it’s truly disappointing that he has struggled to take significant strides in any other aspect of his game.
After hearing a bit about how big data is being used in today’s NBA, I moved on to hear about another research paper, one which featured Grantland’s Kirk Goldsberry as one of its main contributors. The goal of this research was, in a nutshell, to gain the ability to estimate the expected value of a possession at any single point in time based on where the ball and all players are on the court. The correlated single player statistic, called “EPV-added”, evaluates the amount of points per game a player adds to his team relative to the league average as a result of his decision-making ability. This season, the top 5 in the league are Jose Calderon, LeBron James, Dirk Nowitzki, Channing Frye, & Chandler Parsons.
Finally, we arrived at the basketball analytics panel, moderated by Zach Lowe and featured former Toronto Raptors GM Bryan Colangelo, former NBA coach Stan Van Gundy, former player/GM Steve Kerr, Boston Celtics head coach Brad Stevens, and Celtics’ assistant GM Mike Zarren. A lot of great topics were bounced around between these guys, all of whom had fantastic perspectives to bring to the table.
- On shooting 2s vs. 3s: Kerr mentioned the importance of evaluating the consequences of each type of shot that doesn’t go in as opposed to merely using those that do. How do chances for offensive rebounds or fast breaks for opponents change based on where the shot comes from? Van Gundy added his own opinion on this topic, questioning the mental aspect; per Gundy, even though making 1/3 of three pointers is mathematically equivalent to converting 1/2 of two-point attempts, is a player who does the former negatively affected by seeing less of his shots fall?
- Van Gundy went on to question the use of analytics in the NBA at such a micro level. He joked about the numbers saying that the Knicks isolate 17% of the time, but with his eyes he sees them isolating as much as 50% of the time. His main point was not to denounce analytics, but to suggest that using the data in lieu of watching the games is a poor idea. To perform useful analyis, you need visual data to back up numerical data.
- The conversation then turned to player health, from questions about how to prevent injures to concerns in regards to workload. Van Gundy made the point that people discuss why cutting player minutes could be beneficial, but for some reason overlook the fact that when you reduce those good players’ minutes, you lose more. Mike Zarren made light of the fact that teams have so few opportunities to have real “practices” since there are so many games without much time in between.
- Colangelo assessed the kind of chances that there is to break into the league, claiming that there is “mountains of opportunity” from an analytics perspective. Colangelo believes that the market in this area hasn’t even established itself yet, and estimates that an average of between $250-500k is being spent per team on analytics. At this point, Kerr echoed Van Gundy’s earlier sentiments, stressing the importance for those in the industry to know the players and the game along with the numbers.
- On the subject of tanking, the “wheel” concept was brought up (where every team in the NBA picks in every spot in the draft over a 30-year period). The main concern addressed with this method would be if a superstar player in college were to decide to not declare for the draft based on the team picking first in that year. Colangelo admitted to trying to tank a couple of years ago, though not in the sense that he was telling his coaching staff to do so. Van Gundy noted that he is in favor of anything that takes away the intention to lose. Van Gundy on the 76ers: “If you’re putting that roster on the floor, you’re doing everything you can to lose.”
- Brad Stevens said that the difference between NCAA basketball and the NBA is that there are mountains more data to sort through professionally. He he went on to promote Rajon Rondo as one of the players most inclined to study basketball analytics, joking about him maybe being at the conference.
The basketball analytics panel then ended, to be followed by the last one of the day with the topic of building a dynasty. (I should note that it was at this point when I found out that the Pelicans declared Holiday out for the season, and somehow I resisted the urge to kick and scream and throw things.) The panelists were Jonathan Craft of the Patriots and former Bulls and Lakers head coach Phil Jackson (moderated by Jackie MacMullan). Some of the key takeaways:
- Phil Jackson highlighted the importance of having real practices for teams, saying that practices “build character” and give the second unit valuable chances to play against the starters.
- Phil Jackson thinks that the biggest way that being a player helped him as a coach was knowing how players deal with injuries.
- Jackson mentioned Rick Fox first when asked to name the most under-appreciated player that he coached.
After an evening happy hour where I continued to finally attach faces to Twitter handles, day one of the conference came to a close. Day 2 was not as basketball-heavy overall, but did feature an hour long conversation between new NBA commissioner Adam Silver and Malcom Gladwell. Highlights from Day 2 coming soon!
As you watch the gametape match the game actions to your charts....watching video is the BIBLE of the game...:"stats accuse you,but the gametape convicts you"...tape tells the truth for players to see.
New School Rebounding Language: Positioning,Hustle,Conversion...Old School: Win the spot,blockout and get the damn ball for a putback or start the break....
I think this has been posted in the past, but who is the Pels analytics guy and do you have a sense of whether he has much pulls in the organization?
@New City The Pelicans sent one representative to the conference this year: basketball operations analyst Somak Sarkar.
@Jason Quigley every team sent at least one rep.