Dan Devine

Rise of the machines: Will coaches cede control to data analysts?

Dan Devine
Ball Don't Lie

View gallery

.

Tarek Kamil predicts that technologists will one day be just as important to a basketball team's on-court success as coaches are. Not only that, he says, they'll actually be determining in-game strategy and making sideline calls themselves.

And as he sees it, that day's coming sooner than you think. During a presentation on the second day of the 2011 MIT Sloan Sports Analytics Conference, Kamil said he figures it'll arrive within the next 20 years.

Analysts will pore over mountains of data mined from an array of individual player snapshots and game situations to arrive at evidence-based conclusions, according to Kamil, executive director of online strategies for InfoMotion Sports and founder of the popular online sports simulation site WhatIfSports.com. Those conclusions will form the basis for the kind of critical decisions that frequently mean the difference between winning and losing.

Kamil forwarded this idea during a presentation promoting InfoMotion's 94Fifty technology, which is built on a pretty basic premise — if you measure the motion of the basketball as a player handles it, you can get a lot of information about the player doing the handling. The system uses sensor arrays embedded into basketballs to collect a variety of data about how the ball moves when it's dribbled, shot or passed. From there, the system employs advanced algorithms to turn the information into a highly detailed player assessment.

Using the 94Fifty technology, Kamil said, you can learn how much control a player has over the ball — not only the speed of his dribble or the steadiness of his handle, but also what kind of spin he puts on the ball when he shoots it, how long it takes him to get different types of shots off in different situations, the angle and arc with which he shoots it, and more.

"A coach can tell me that [a player] needs to work on his left hand," Kamil said. "[But] we can tell you that his right hand is 14 percent more dominant than his left."

That degree of granularity is more than just interesting, according to Kamil — it's necessary.

"How do you get better if you don't know where you are today?" he asked.

View gallery

.
In addition to evaluation and player development purposes, Kamil said, advances in data collection like the 94Fifty system have massive potential for tightening up teams' game strategies.

Take substitution patterns, for example. By comparing data culled from early in a game, when a player is fresh, with information recorded near the end of an extended period of play, Kamil claims the system can show just how much a player's performance declines due to fatigue, helping to pinpoint the optimal time to yank a gassed starter in favor of a rested backup.

Ditto for players' effectiveness in relation to one another and in specific circumstances, which can aid in choosing the best possible lineup combination for protecting a sizable lead or erasing a daunting deficit.

Transferring the responsibility for such game-time decisions to analysts would allow coaches to focus more on the human element of their profession — teaching players the finer points of being a professional, emphasizing organizational structure and discipline, and managing relationships within the locker room. No longer would coaches serve as sideline tacticians; instead, they'd become something more like experienced, extremely hands-on mentors.

This is where your inner traditionalist starts calling B.S. (If it didn't start hollering after the first paragraph, that is.) To be honest, my immediate response at the moment Kamil made his pronouncement was the same as Zach Harper's — it's really, really hard to envision coaches actually signing off on this arrangement.

Sure, technologists and analysts are experts in technology and analysis, but basketball coaches, at least in theory, are experts in basketball. Shouldn't that matter? Moreover, many coaches are notoriously micromanaging control freaks that live in constant fear of losing their jobs. You think they're going to just willingly substitute some bean counter's calls for their own judgment when they're the ones getting raked over the coals following every loss?

During Friday's "Gut vs. Data: How Do Coaches Make Decisions?" panel, ex-Texas Teach head football coach Mike Leach stayed pretty close to the traditional line. He said that no amount of analysis can replace players having a clear mind, emphasizing the importance of purposeful, repeated practice of game situations to get players ready for high-leverage moments. In describing his coaching style, he even used a Sharon Stone quote from a 1992 interview with Playboy: "No guts, no glory, right?" (It is unlikely that she was the person who coined that particular phrase.)

Of course, football and basketball are different. Plus, y'know, Leach is an ex-coach whose decision-making, both on-field and off, has famously come under quite a bit of scrutiny. The "ex" is relevant here; Kamil thinks the quest for job security will play a large role in the increased adoption of analytics by coaches.

The bet Kamil's making is that teams increasing their emphasis on analytics and making more intelligent use of data will see meaningful improvement on the court — not a wholly unreasonable position, considering pro-analytics organizations like Daryl Morey's Houston Rockets, Mark Cuban's Dallas Mavericks, R.C. Buford's San Antonio Spurs, Sam Presti's Oklahoma City Thunder and Danny Ainge's Boston Celtics are all either enjoying success now or have in the recent past.

As more owners and executives see that this stuff works, Kamil argues, coaches won't have a choice. If they're not willing to elevate empirical fact over the healthy hunches for which they've long opted, they'll soon find themselves looking at the want ads. That makes sense — after all, the golden rule in business has always been that the guy who has the gold makes the rule — and fits well into the narrative of evolution in a sport that was played with a peach basket 120 years ago, Chuck Taylor All-Stars 55 years ago, and short-shorts 20 years ago.

The challenge for analytical types, then, appears to be getting the top lines of the organizational chart to buy what they're selling. The process has already started — Zach Lowe noted at The Point Forward that about half the teams in the NBA sent representatives to Sloan, and Celtics front office stat guru Mike Zarren said during a panel that 20 of the league's 30 teams are pursuing analytics on some level. It may well be just a matter of time before coaches are trading in their clipboards for tablets and having the right out-of-bounds play dictated to them by a program that has simulated that exact scenario 10,000 times, rather than drawn up by them based on an amorphous sense for what just might work at the moment.

It'll feel weird at first, but we'll get used to it. We always do.

View Comments