The NBA will officially announce Thursday what Grantland reported two weeks ago — that the league will pay for the installation of data-tracking cameras, and the attached software, at all 29 NBA arenas (the Clippers and Lakers share an arena). This is a sort of endgame for STATS, proprietor of the SportVU camera technology, which entered the league in the 2010-11 season with a half-dozen eager subscribing teams. I’ve written about the technology several times, so I won’t go deep into the basics here. Suffice it to say the cameras track the movement of every object on the court — players, referees, the ball — several times per second, providing a new path to answering questions small and grand.
The potential impact on our understanding of the game, of its flow and X's-and-O's, is fascinating. The NBA has become the first major U.S. sports league, and perhaps the first in the world, to invest this heavily in motion-tracking. But the cameras will touch on lots of other areas of profound importance to basketball’s future that have gotten short shrift amid the hoops-related curiosity.
Half the league’s teams had already purchased the camera systems at about $100,000 per year. The league's move to pay for the remaining 15 teams caught a lot of folks close to the process by surprise. It won’t tout it, but one reason the league acted fast was to immediately enhance its ability to monitor referees — always a touchy subject. The cameras represent the most precise way to grade the three on-court officials based on how consistently and early they get into the league’s three set positions — called “lead,” “slot,” and “trail” — and whether they make appropriate calls from those positions based on their exact sight lines. This is the next stage in seeing which officials are the best, and thus deserving of high-stakes assignments, and in quantifying that in ways that are hard to dispute.
“We will use whatever data and means we can to improve our referees,” says Steve Hellmuth, the NBA’s executive vice president of operations and technology. “The refs haven’t been tracked before. Now for the first time, they will be.”
The league has already started using the cameras to check on the enforcement of defensive three-second violations out of concern that defensive players routinely break the rule by lingering in the lane too long. (The results of said studies are inconclusive so far, say several sources familiar with the inquiry.) One catch: It’s unclear if the league will share any referee-related data with the subscribing teams. “I don’t think we’ve thought about that at this point,” Hellmuth says.
The league has long hoarded data on which individual referees make particular foul calls, something you won’t find in the play-by-play and a key piece of information a few teams have paid outside consultants to track.
The plan for now appears to be for the league to keep the camera-related referee data to itself, a move that will not please teams.
Fitness and Practice
Teams can pay up to $40,000 extra to purchase (among other goodies) software that helps track a player’s physical exertion. The in-game cameras represent one piece of that. They can tell you how fast a player runs, how often he accelerates on cuts, how often those accelerations end with him reaching top speed, and the height of a player’s release point on jump shots. Some players recovering from injury, including Ricky Rubio last season, have taken significant game time to get back to their previous speed and fitness baselines. And an injury to one star, Manu Ginobili early in the 2011-12 season, resulted in the other San Antonio starters exerting more physical effort with a standstill shooter (Danny Green) in Ginobili’s place.
The other pieces, and perhaps the most important ones in determining a player’s condition, come outside those 82 games and require the use of other forms of technology: sleep and heart-rate monitors, GPS devices and accelerometers players can wear during practice, and the careful tracking of weightlifting, diet, and other day-to-day stuff. Put all that data together, and you can get a fairly complete picture of a player’s condition, and of how indicators of his condition — running speed, jumping ability, etc. — change over the course of a season. “This is where you can start to measure fatigue,” says Brian Kopp, executive vice-president at STATS.
A revealing nugget: Teams really want the SportVU cameras to monitor their practices, Kopp says. That’s difficult, since most teams practice somewhere other than their game arenas. Some coaches and GMs might want the practice data simply to check on which players work hard, and which loaf.
But others will want it to change the very concept of practice. How much practice time do teams really need? And how taxing should those practices be? How should that change during the season? There are higher-ups around the league who are ready to radically rethink these things, provided the next-level data indicates they should.
The Box Score
It has been more or less the same since the formation of the NBA, with new statistics sprinkled in as the league decided to record them. But it’s outdated, and sites like Basketball-Reference have long provided more advanced box scores, with numbers like pace, offensive and defensive efficiency, rebound rates, and others.
Hellmuth and Kopp think the camera data could bring the next step in the reinvention of the box score, and that the NBA can popularize that reinvention by putting new numbers in front of fans via NBA.com, NBA TV, and mobile device applications. Building a new box score will take years, maybe a decade, and nobody really knows exactly what stats should be in there. Kopp and Hellmuth mentioned hockey assists, total distance run, and even rebound chances — the number of times a player was within, say, three feet of an available rebound. (The Rockets signed Carlos Delfino before last season in part because the camera data revealed he grabbed an unusually large percentage of rebounds that fell near him. That’s a nice extra skill for a wing shooter.)
The box score is only one component in improving the way fans understand the game, Hellmuth says. He envisions a day when the cameras will be able to educate fans on the spatial aspects of defense — where players are supposed to be against a particular opponent running a particular offensive set. (The Raptors are furthest along in understanding this. Implementing it is another matter.) “We can really advance the fans’ understanding of our game,” he says.
The Best We Can Do
The league could theoretically purchase military-grade motion technology that would cost tens of millions of dollars for each arena. They could also insert microchip technology into the ball itself that would provide a finer portrait of its path around the court and in the air. Those things aren’t happening, and the players have made it especially clear they will revolt against any change to the ball itself — just as they did in 2006, when the league briefly switched to a synthetic ball. “The players are simply too sensitive to any changes,” Hellmuth says.
The cameras aren’t perfect. They don’t indicate which direction a player is facing, though teams can generally figure that out. And though they can measure the flight path of the ball, they cannot yet measure the vertical leap of each player.
But they’re damn good, they’re affordable, and they’ll get better — both on their own, and in conjunction with other technologies.
Only teams and the league are set to have this data for now, and the teams, as stated above, get only the player data — not the referee stuff. So imagine a player entering the final year of his rookie-scale contract and his agent beginning contract talks only to hear a team official open with something like, “Our camera data shows you really don’t hustle in the fourth quarter. Your running speed slows down. You just stand around instead of going for rebounds. These are some of the reasons we are offering you only $7 million per year.”
Wouldn’t that agent want to at least cross-check that data, to make sure it’s not B.S.? The players union has already started the fight for access to that data. “All we want is to make sure access is available,” says Ron Klempner, the union’s executive director. “If teams are forming impressions about players that players are not in position to defend, we want to make sure everyone is operating on an even scale.”
If teams really want to get something out of this data, they’re going to need to hire different sorts of people. STATS provides every team sophisticated reports and access to a database of information, but the smartest teams will find ways to mine the raw material the cameras spit out before STATS organizes it — basically, a Byzantine set of coordinates that is borderline incomprehensible to laypeople — and learn things they don’t have to share with anyone else. They could draw conclusions about anything: the best way to approach offensive rebounding without sacrificing transition defense; substitution patterns related to foul trouble; ideal shot selection models against particular teams; objectively better ways to defend the pick-and-roll; and hard-to-discover tendencies about individual players.
Teams will indirectly share some of those discoveries by how they play on the court, which players they sign, and how they approach trades. But every competitive advantage matters, and to get them, teams will have to invest in different sort of analytics people — sophisticated computer programmers who also understand basketball — and stats personnel with a better knowledge of computing, hoops, and the appropriate questions to ask of the data. “It can’t be just, ‘Let’s hire some smart Harvard dude we can wear out for 15 hours a day before he realizes he can make a lot more money on Wall Street,’” Kopp says. “Teams are going to have to pay for good talent. There are a lot of spreadsheet analysts with teams now. But teams should be hiring more programmers.”
Some teams will get almost nothing from this data over the next half-decade; some aren’t all that interested, in part due to skepticism that it will ever have any concrete impact on coaching strategies or player acquisitions.
Some of the teams that have had the data have done almost nothing with it. In some cases, that stems from a lack of curiosity and/or staffing issues that make diving into this giant pile of information look like an impossible task amid a dozen other more immediately pressing jobs. But some teams haven’t hired personnel capable of taking the information and making it into something more.
The Looming TV and Digital Rights Deals
It’s a footnote to a very expensive story, but if this data becomes popular quickly, the league may be able to introduce it as a small element in its TV rights windfall. Under the league’s current national TV deal, set to expire after the 2015-16 season, ESPN and TNT pay $930 million combined annually to broadcast games. The players get roughly half that money under the terms of the new CBA, down from 57 percent under the old one, and the league splits its portion among the 30 teams.
The league signed that deal during a relatively low point in popularity and ratings (in 2007), and the price for broadcast rights to live sporting events has ballooned since, as sports has emerged as perhaps the only DVR-proof form of TV entertainment. The league’s new TV deal is going to be a bonanza, especially with Fox Sports 1 rising as a third major bidder. The league might be able rake in a little extra cash by offering SportVU data as an exclusive nugget to either a bidding network or a mobile company that might design an application around the data for hardcore fans.
Regardless, the cameras might seem like a wonky, small development that doesn’t matter much. It may end up that way. But I suspect otherwise. The league has made a commitment here, and it has opened up a lot of new possibilities.