The growing move to measure everything, including basketball

This time it is from the NBA:

The league on Thursday announced plans to install sophisticated tracking cameras, known as the SportVu system, in every arena for the coming season, creating an unprecedented treasure trove of data about virtually every wrinkle of the game. SportVu, developed by Stats LLC, records data points for all 10 players, the three referees and the ball, every 30th of a second, measuring speed, distance, player separation and ball possession. Every step, every dribble, every pass, every shot, every rebound — really, every movement — will be recorded, coded and categorized. … The N.B.A. is the first major professional sports league in the United States to fully adopt the SportVu system. It will have other implications for the league, far beyond the playbook and the box score. Not everyone might welcome the change. General managers will surely exploit the more sophisticated statistics when negotiating contracts with player agents. Not all assists, points and rebounds are created equal — and teams will soon be able to demonstrate that vividly. Referees are also tracked by SportVu, which means the league will have yet another tool to analyze every call, non-call and missed call as it ranks its officials. Those rankings help determine which referees are chosen for playoff assignments and the finals.

There is yet further information here.  One prediction is simply that players will pass the ball more, even when it does not result in an assist.  Team defense will improve too.  Furthermore some injuries may be partially forecastable and thus preventable.  If applied at the college level, the efficiency of the draft will improve and this will help restore competitive parity.  Truly injured or otherwise disadvantaged players may lose some insurance value, since it may be clear earlier on they are not going to recover or improve.


"General managers will surely exploit the more sophisticated statistics when negotiating contracts with player agents". It seems to me that most professional basketball players are grouped narrowly around the mean. Except, of course, for those who stretch the tail of the distribution out the right. The negotiating power may be with the players rather than with the general manager, no?

I had similar thoughts. Unless there is some sort of asymmetry, for example, if only the GM of the player's team has access to the stats, then I would think that more transparency would favor the better players and disfavor the worse ones. Similarly, for the better refs vs. the worse refs. Actually, "favor" is the wrong term since better ability to evaluate performance actually increases meritocracy.

One earns premiums for being differentiated from one's competition. That's true for both firms and individuals. These stats allow players, the good ones at least, to differentiate themselves (even more than they already do with current stats), making them less replaceable by other players, thus able to command higher premiums.

Arguably, one of the reasons that professional athletes earn such astronomical salaries is because, even without these new stats, all of their team's competitors can observe their play and, thus, have a pretty good idea who the best players are. Such transparency is what leads to fierce competition in the free agent market. That's much different from most other professions, where typically a firm has a much better understanding than its competitors of who its key employees really are. That information asymmetry undoubtedly allows the firm to capture more of its key employees' value-add than would otherwise be the case with greater transparency.

So what about CEO salaries?

CEOs are the ones with the most external visibility and are not viewed as interchangeable. (Some might argue that they should be viewed as interchangeable, but they're not viewed that way.) In fact, like athletes, most would-be CEOs make $0 because no one will hire them to be CEO. The ones that are hired earn premiums for their perceived uniqueness and differentiation.

That perceived uniqueness is related to the fact that there is very little external visibility/transparency into their underlings. Thus, when a firm is successful, the CEO gets credit. Professional athletics, though, with their high transparency are quite different. Athletes (labor) make the most money, then coaches (lower management), then front-office executives. Why? Because everyone can see that (particular individual) athletes are most responsible for the team's success, then (particular) coaches, then (particular) front-office executives.

This will affect the college game too. I predict that well-heeled college programs will soon join the analytics avalanche. Others will ultimately do so out of necessity to keep up, especially if the price of the system comes down over time.

I could be wrong, but I think none of this matters to championships (although it matters at the boring level of who is the third or fourth best college team east of the Mississippi and west of the eastern time zone ( etc.). 2 or 3 times a year some little baby is born who has the right genetics and cares so much about b-ball (or number theory, or blues, of the Western Canon, or the difference between long-dead Austrian theorists and some idealized version of a logic-scholar-Keynes avatar) that he is gonna coach his way, or point guard his way, or small forward his way, to winning everything available, at least once and ideally several times (or do well in the other fields mentioned). ... There are approximately 320 x 100 iq points in this country and I for one am tired of having so many of those points wasted on festivals of frivolity. Lets make the people who cure annoying diseases and fix general human problems richer than the sports "aficionados"

I've read Bill James avidly since 1985, but the real world changes brought about by advanced sports statistics have turned out to be surprisingly marginal. That's because, as Yogi Berra said, you can observe a lot just by watching.

One measurement every 30th of a second? Sounds like there's more to this than stats. This is the kind of data video game programmers would use to recreate/simulate virtual games.

I don't think much will change. For one, half of teams already had this last season. Also, there were a number of "marquee" presentations using this data at the Sloan Sports Analytics conference in March--everyone was really excited about it, but none of them were particularly impressive. It may take a while before any front office can exploit the new data.

If anything, this may give more of an edge to well-managed teams. Many basketball GMs are not close to optimal using existing box-score data and manage to screw up surprisingly simple draft and salary cap-related decisions. Now they're going to try and use this complicated data source when they can't even apply a simple rule like "ignore volume scoring and focus on efficiency".

How about the intangibles? The guy who reads situations and adapts on the fly but doesn't have particularly good athletic skills? The guy who knows where to be on the floor, the guy who inspires his teammates or calms them down. We start trying to quantify everything and sure enough we're going to think only quantifiable skills are necessary and sufficient and next thing you know we're drafting athletes with no fundamentals. Wait...

This reminds me of an article I read about McNamara losing us the Vietnam War by focusing on enemy casualties as a metric for success.

These are being factored in today. See: Shane Battier. Not the most athletic, but was highly valued because of certain attributes (good shooter, forced defenses to be more spaced, always in the right spot at the right time, etc.)

Reading the situation and adapting on the fly is not really an intangible. Positioning? Not an intangible. Giving good talks to his teammates might not be currently measured, but it sure can be: Don't track what they say, just track that something was said, and then look at results.

And really, if a locker room cancer really affects a team, you'll be able to tell by how much better his teammates get when they leave.

Clearly we've picked all the low hanging fruit in the market in everything about stories about 3D printed, mood affiliated driverless cars in Haiti as told by Deirdre McCloskey. Let's speculate about what Paul Krugman thinks about this with a scathing rebuttal by prior approval about overpaid tenured professors at GMU funded by the Koch Brothers.

Wonder if anyone has applied facial / body language analysis to chess. Could one detect an opponent with a key move up his sleeve or who has just discovered his blunder or an exciting option?

To the extant that there is certainty of player excellence, it should clarify negotiations, not advantage one side or the other. It takes risk out of paying a high price (GM) & asking for a high price (player).

Comments for this post are closed