If you want to see how the NBA imagines its future, don’t bother watching the court. Instead, squint into the rafters and try to find the six tiny cameras that are part of the SportVU tracking system being implemented league-wide this season. Derived from Israeli missile-tracking technology, SportVU records a player’s every dribble, pass, and off-ball movement 72,000 times a contest – that’s 25 times per second. The big idea: Motion-capture software sophisticated enough to defend against an incoming Scud should be able to distinguish between a lethal Stephen Curry three-point bomb and just another errant Monta Ellis long-range brick.
These eyes in the sky will provide reams of real-time data instantly crunchable into an array of advanced statistics. Tracking every pass – including free-throw assists (passes that led to a player being fouled), secondary or “hockey” assists (passes to the passer that got credit for an assist), and potential assists (passes that would have led to assists if missed shots had been made) – will determine who’s actually generating the most offense and setting up teammates for the best shots. Being able to intricately follow players and the ball will reveal who’s securing the most rebounds in traffic and who’s creating the most rebound chances. Metrics related to protecting the rim (how a defender near the basket affects both shot selection and field-goal percentage) will offer an assessment of defensive impact insufficiently detailed by such fuzzy statistics as blocked shots. “Up until this point, they haven’t been able to measure these adequately,” says Brian Kopp of Stats, the pioneering company that purchased SportVU five years ago and adapted it for use in the NBA. (It is also used in soccer and could eventually come to the NFL.) “Hopefully some of these new statistics become part of the language of basketball.”
The NBA’s embrace of SportVU is just the latest milestone in sports’ numbers revolution. It’s been 10 years since Michael Lewis’ bestseller ‘Moneyball’ injected the once-arcane concept of advanced stats into the mainstream. Though sabermetrics (which its inventor, writer and statistician Bill James, defined as “the search for objective knowledge about baseball”) and analytics (the use of data to discover patterns and predict outcomes) rode baseball’s bench for decades, the success in the 2000s of GMs like the Oakland A’s Billy Beane and the Boston Red Sox’s Theo Epstein eventually led to a proliferation of MLB front-office number crunchers. Despite some resistance, the other major sports are finally beginning to follow suit. An NFL team’s search for a winning formula is now as likely to include elaborate algorithms as hard-nosed coaches. In hockey, so-called player usage charts display how well a player is utilized in various offensive and defensive situations. Heat-map graphics show the range of player movement on the soccer pitch. Blogs, fantasy leagues, and online gambling all thrive on the new numbers. Sports fandom may be intrinsically subjective, and purists may complain that all this data reduces the intangible poetry of their beloved games to mere trigonometry, but analytics are here to stay.
The ramifications will potentially be felt far and wide, from the front office (including player evaluations, roster acquisition, and contract incentives) to the bench, where coaching strategy has moved from scrawling x’s and o’s on a whiteboard to plotting X/Y data on a laptop. Even the training staff and strength coaches will be affected: Now they can monitor just how much distance a player has covered on the court, or calculate the force being exerted on recovering knees. “It’s exciting and also a little daunting,” says Memphis Grizzlies vice president of basketball operations John Hollinger. “It’s going to unleash a tsunami of data, and you have to be able to process all that information or you won’t be able to get anything out of it.”
To stubborn naysayers or those who simply find it confusing, Daryl Morey, GM of the Houston Rockets and the first analytics adherent to helm a non-baseball American sports franchise, bluntly defends the use of data: “It works. The bottom line is that sports are about decision-making, and the science shows over and over that if you use objective data and analysis in decision-making, you make better decisions.”
Morey has been the de facto leader of the movement. Instead of the typical old-guard NBA GM background as a player or coach, Morey has a computer science degree from Northwestern; as an undergrad he worked part-time at Stats, where he became the first person to apply Bill James’ “Pythagorean expectation formula” to basketball. After getting an MBA from MIT, he worked at a business consulting firm, then as a Boston Celtics executive, before Rockets owner Les Alexander (a former Wall Street trader) tapped him to be GM. While with the Celtics, Morey co-founded the MIT Sloan Sports Analytics Conference in 2007, which has grown from a gathering of 175 people to an entire weekend every March with 2,700 attendees (and hundreds more turned away) presenting papers like “Acceleration in the NBA: Towards an Algorithmic Taxonomy of Basketball Plays” and “Going for Three: Predicting the Likelihood of Field Goal Success with Logistic Regression.”
The roots of basketball analytics actually date back to the 1960s, when University of North Carolina coach Dean Smith charted such non-box-score factors as rebound attempts and “hustle points” (baskets scored off gritty effort). But the modern era’s numbers boom began in the early 2000s, in the same grassroots fashion as baseball’s, with posters on message boards like the Association for Professional Basketball Research expanding on the innovations of obscure statisticians like Bob Bellotti and Dean Oliver. Many NBA teams now have analytics experts on staff – for example, in 2012 the Grizzlies snapped up Hollinger, then an ESPN writer, who had developed the Player Efficiency Rating, an attempt to assess all of a player’s offensive and defensive contributions in a single number. Hollinger’s remarkable career arc – from proto-blogger to columnist to an executive with roster input – is emblematic of the new age. “I think a lot of people on this side have embraced the word geek as a badge of pride,” he says.
Last off-season, the Boston Celtics named former Butler University coach Brad Stevens, an outspoken analytics devotee, to head one of the NBA’s most hallowed franchises – though Stevens feels that all the talk of advanced stats is somewhat overblown. “Coaches have always used whatever information they had; the difference is that now we all have way more information,” says Stevens. “You’re always looking for any little edge you can get. From a coaching standpoint, you try to put your guys in positions where they can best be successful, and a lot of those numbers help you do that.”
Some coaches, however, have proved more reluctant to surf the wave of new facts and figures. Shortly after the Grizzlies’ new stats-friendly ownership hired Hollinger, coach Lionel Hollins publicly criticized analytics in a radio interview. Despite the fact that Hollins led the Grizzlies to a franchise-record 56 wins and its first-ever berth in the Western Conference finals, the team decided not to renew his contract.
“I still trust the gym, practice, and games more than I do statistics,” says ESPN analyst and former NBA coach George Karl. “But I do think there’s space in the game where analytics can help a coach, team, or organization figure out what decisions they should make.” Citing the 11 coaches who lost their jobs this past off-season (including himself, let go by the Denver Nuggets), he sees a more ominous role that numbers can play – one that pressures coaches not just to win, but to win a specific way. “You could probably use stats to prove that a coach who won 56 or 57 games still provides reason for a change. Winning was always the security blanket of coaching, but now it’s winning plus you must follow the philosophy the organization wants. It’s kind of confusing to coaches. It feels like sometimes you’re drowning when you’re winning games and you’re being told you’re not doing the right thing.”
Of course the most important constituency that needs convincing may be the folks being rendered into data: the players. “I try to use analytics mostly on the defensive end,” says Miami Heat forward Shane Battier, who requests info on opponents before games. “If I’m covering Kobe Bryant and I know that if I send him to his left hand versus his right, I’m shaving 50 basis points off his efficiency, over the long run that’s going to help me.” Without his use of numbers, says Battier, “I don’t think I would be playing at a high level at 35. I can’t jump as high as I used to, I’m not as fast, but I understand the game of basketball. I know what gets me in trouble and I have the data to back me up, and it has enabled me to carve out a unique niche in the NBA.”
But Battier remains an outlier among athletes. Only a “handful” of NBA players, he says, “maybe one or two on every team, really understand how to apply the numbers.” Battier feels the problem may be intrinsic to the sport. “There is always going to be a conflict with how practical analytic theory can be in a game because the game is so fast-paced,” says Battier. “As much as you try to make it academic, it’s an instinctual game at its heart, a primal game.”
Still, that hasn’t stopped him from attempting to be a transitional figure in the sport. “Like it or not, big data is driving the world,” says Battier, “from the stock market to agriculture to entertainment and sports. The more data you have, the better served you are. That’s why I think teams are investing in analytics personnel – and if they don’t, they’re going to get left behind.”