NBA Analytics Royalty Steps Out at SSAC 2017

(Photo cred: @SloanSportsConf Twitter)

What are the most common misconceptions about the three-point revolution? How do you construct a team around a superstar? Which NBA teams run the most precise sets? What are the secrets to negotiating the right deal with a big-name free agent? How do players, coaches, and general managers view the future of basketball analytics?

These questions and many others took to the stage at the 2017 Sloan Sports Analytics Conference (SSAC), the annual industry trade show for the burgeoning intersection of sports, technology, and data, on March 3-4. This year, more than 3,000 participants packed the Hynes Convention Center in the center of Boston, MA to hear the latest and greatest from league executives, former players and coaches, current and former general managers (at least six of them, by my count), technologists, journalists, and researchers. While the conference isn’t exclusively focused on the NBA, it has become the largest single focus of the conference over the past few years, thanks to the growing popularity of the league, the widespread availability of exoteric advanced statistics, as well as new types of data (such as SportVU) yielding interesting new research topics.

The two days at SSAC are packed with insightful panel discussions, research presentations, and interviews. (They’ve even got classes on statistical programming, résumé workshops for those looking to break into sports analytics professionally, which is a lot of the attendee base, and this year introduced eSports competitions.) In fact, the weekend is so chock-full of action that I felt mentally exhausted by the last few sessions. I won’t try to summarize everything, but there were a few themes and topics especially pertinent to the NBA and to the Utah Jazz that I’ll share here.

• A few of the panels brought up the current lack of good analytics for the influence of coaching. In short, beyond wins and losses, we don’t have good, hard data to help us understand which coaches are better than others. Dean Oliver, whom we might call the godfather of NBA analytics, suggested that the best we can really do right now is look at a player’s trajectory before working with a particular coach and compare it to his trajectory while/after working with that coach. To me, that seemed fraught with noise, because players make leaps for many reasons, not all of them related directly to coaching. Part of the problem is that, unlike players on the floor, we don’t have a lot of visibility into everything coaches do: how they motivate players, what happens in practice, etc. We do have a lot of data on lineups and rotations, and that could yield some insight if applied in the right way. I think the general consensus among those who brought this up at the conference was that this is an area rich in opportunity for future research.
• Specialization among young athletes came up a few times. There is plenty of evidence suggesting that specializing—picking one sport and sticking with it—is not good for kids. Parents, don’t force your kids to play just one sport exclusively. I know it might sound like a great idea if you want your child to have the opportunity to play at the next level—the now-debunked 10,000 hours theory or whatever—but it leads to injury and wear on the specific muscles and ligaments used in that sport, it causes burnout and emotional issues, and Shane Battier felt strongly that it robs kids of their childhood. (Battier was first recruited by University of Minnesota as a 12-year-old.) Whenever this came up, I thought about Gordon Hayward. What if he had stuck solely to tennis? Matt Harpring is also a notable former two-sport star who would not have had a lengthy NBA career if he had played only football.
• The most discussed new research at SSAC 2017 was a paper presented by Andy Miller called “Possession Sketches: Mapping NBA Strategies.” Miller showed how they were able to use player tracking data (SportVU) to find the “semantic structure,” for example, a four-man weave resulting in a Klay Thompson three-pointer, of every offensive possession so that coaches and video teams could easily find the plays they want to analyze. Imagine a search engine that would let you find all plays of a specific type, involving specific players, with a specific outcome, and then return the video clip of every one of those plays. If good offense beats good defense, then you have to stop good offense. Being able to quickly find and analyze every Spurs hammer set is key to that. Miller’s research also included 3D mapping of how precisely teams executed specific sets. I would look at this in part as a team’s ability to impose its offensive will on opponents; which teams are able to get to their spots and run their plays most consistently? And, of course, the point is to teach your defense how to disrupt what a team is trying to do as soon as you recognize it.
• Perhaps the most anticipated session of the conference was “Silver Asks Silver,” where Nate Silver of fivethirtyeight.com interviewed NBA commissioner Adam Silver. Both Silvers are engaging conversationalists, but the commish always comes across so genuine that I find myself wanting to ask him life advice. One of the major topics that A. Silver discussed was competitive balance between large-market and small-market teams. His view is that the internationalization of the game, in large part due to streaming and social media, will lessen the advantage that large-market teams have in signing and retaining talent because players can build their brands no matter where they play. In the past, you were really isolated playing in Utah or Orlando; if you weren’t on national TV, fans couldn’t see you play. Now, not only can fans all over the world watch Rudy Gobert and George Hill, but they can follow the players on Twitter and Instagram and Snapchat, letting fans have a closer relationship with their favorite stars who play in small markets. Personally, I think this is a bit reductive; there are plenty of reasons why players might want to play in a large market, and building their brand is only one motivation. But he is right that the ability for fans to “connect” emotionally with stars is a huge driver of league growth.
• A few other quick-hitters from Adam Silver:
  • Expansion into Europe is unlikely due to travel as well as European economies, but Mexico City is a real possibility.
  • Chris Paul approached Adam Silver after the All-Star Game last month: “We need to fix this.”
  • Silver does not mind players and coaches taking political stances. He prefers that they represent NBA values in their stances, and couch things in those terms.
  • The commissioner talked about the salary cap a bit, noting that a hard cap with no max on player contracts is unlikely to happen. Someone would spend 80% of their cap space on one superstar and surround him with minimum contracts, which is bad for quality of play. When we talk about the CBA and media deals as fans, we often focus on the max-money guys and how these changes will impact our favorite star players. Silver reminded us that the players’ union is actively working on driving more money to the “rank-and-file” (his words) players, which goes against this idea of a hard cap with no max on individual player deals.
• Seth Partnow, editor emeritus of Nylon Calculus now working in the Milwaukee Bucks front office, gave a talk on data around the three-point revolution in the NBA. He pointed out that the NBA isn’t “becoming a jump-shooting league,” because it has always been exactly that. Last season, 55% of shots were jumpers. In 1999-2000, 55% of shots were jumpers. The percentage of those jumpers that are three-pointers is increasing. He debunked some other myths: the data shows that long jumpers do not actually lead to more fast breaks going the other way, because they also mean better floor spacing and balance; and the mid-range game has diminished, but only for assisted mid-range shots. Frequency of unassisted mid-range shots hasn’t changed substantially over the past several seasons. The Nylon Calculus crew has always railed against 3FG% Allowed as a defensive metric, and Partnow pointed out that we don’t really know from existing data what constitutes a “good contest” of a three, as opposed to a “bad contest” of a three. The best defenders are able to prevent the three point shot from going up at all.
• Evan Wasch, SVP of Basketball Strategy and Analytics, is quickly becoming my favorite perennial speaker at SSAC. Last year, Wasch gave a great talk on the new method of creating the regular season schedule. This year, his presentation was entitled “Anatomy of the NBA Ecosystem,” and focused on how the league office thinks about the relationships between the various stages of the NBA season, from the Lottery (the unofficial “start” of the following season) to the crowning of a champion, and everything in between. Fans often think of potential changes in a vacuum, as impacting only one aspect of the NBA season, when in reality there is a domino or trickle-down effect that Wasch’s team considers. For example, some fans (particularly Western Conference fans, recently) wonder why playoff seeding happens by conference, rather than simply taking the top 16 teams across the entire league. First, Wasch pointed out that, historically, this change would only affect one team per season, so the discrepancy between conferences isn’t as great as fans like to believe. Then he explained that abolishing conferences would result in far more travel, which in turn results in a lower quality of competition on the floor. Currently, 55% of playoff matchups occur between teams in the same time zone; in a 16-seed playoff, that would drop to 30%, and 18% of playoff matchups would span three time zones (e.g. the Clippers playing a first-round matchup against the Hawks). Players would get tired, and the end result would be a worse product on the floor. One thing that the league did change to improve the fan experience was extend the “last two minutes of the fourth quarter” rule against hacking bad free throw shooters so that it now applies at the end of every quarter. As a result, hacking is down 70% this season. Wasch’s point is that they felt comfortable making this change to improve the enjoyability of the game only after analyzing the Pythagorean expectation of teams employing the hacking strategy, and found that across the entire league, hacking only shifted 1-2 wins last season. Fascinating to get a bit of behind-the-scenes insight into how the NBA approaches potential rule changes and structural changes, from the guy whose team does the actual analysis.
• Last thing: Bob Myers, general manager of the Golden State Warriors, joined David Griffin (GM, Cleveland Cavaliers), Masai Ujiri (President, Toronto Raptors), and Luis Scola (who insists he is not retired) on a discussion panel about “building around a superstar.” There were many great insights given into how these GMs think about the roles of their superstars as the nucleus of their teams, but I liked best what Bob Myers said: in his view, superstars are motivated more by aversion to losing than by the joy of winning. I can think of a few current and former Jazz men who seem to fit that mold.

This really is just the tip of the iceberg, believe it or not. I felt as if this year’s conference were a little fresher than the previous year, when we heard a lot of the same old topics from the same old people.

If you’re passionate about sports analytics, or even if you just enjoy hearing from people who have spent their careers in and around the sports you love, start saving now for SSAC 2018. The conference is open to anyone, and you can sign up at sloansportsconference.com to receive email notifications when tickets go on sale, usually in the fall. Conference fees amount to around $600, and it is always held in Boston, so plan accordingly for airfare and accommodations. It’ll give you far more than a year’s worth of topics to think about, and even to research on your own. (And yes, you can absolutely meet and get a photo with many of the speakers.) I’ll hope to see more of Jazz Nation there in 2018.