How the Front-Office Analyst Took Over the NBA

Despite the increasingly frantic protestations of people like Charles Barkley, analytics—and the analysts that come with it—has fully arrived in today's NBA. Sure, there hasn't been a single seminal moment to herald the movement, as there was for baseball with the 2002 publication of Michael Lewis's Moneyball, and nobody yet (at least, not to VICE Sports' knowledge) has tried to get Brad Pitt to play Houston Rockets general manager Daryl Morey in a major feature film, but the change in how NBA teams go about their business has been no less evident for its relative lack of publicity.

And here's the thing: it has mattered. NBA teams employ way more analysts today than they did even a few years ago. The teams that employ more analysts generally win more games, and the teams that got on the train early—adoption before 2012 seems to be key—have won the most of all. We know. We ran the numbers.

Let's start with the first claim: that NBA teams employ way more analysts than they did even a few years ago. Although there's some ambiguity to who counts as an "analyst"—some teams have staff who split duties between analytics and traditional scouting, for example; some employ consultants rather than hire in-house staff, and some GMs, Morey included, are so involved in their team's number crunching that they should count as analytics staff—determining this was the easiest part of the process. We began by scraping data from RealGM's lists of NBA front-office staff, which gave us an initial list of pretty much everyone working in an NBA front office in any capacity.

To narrow our pool a little bit, we isolated those individuals whose titles included words like "analyst" and "analysis," and also personally checked the entire list to make sure we weren't leaving anyone out on the basis of a title we'd somehow missed. We also did a bit of digging to make sure we weren't missing someone obvious—Morey with the Rockets, for example, or Taylor Snarr of the Utah Jazz—or including someone who had since left the team. It wasn't a perfect system (we decided, for example, to ignore the use of those pesky contractors and consultants, as we couldn't obtain reliable data for these individuals), but we're still pretty confident that we now have the most complete list, anywhere, of the NBA's current analytics crowd and, more important, a clear understanding of when they started working for their teams.

The rise, when viewed from 30,000 feet, is stunning: In 2008—the first year we included in our sample, as there were few analytics professionals working in the NBA before then—just five teams employed analysts (and in most cases it was just one analyst). That number steadily rose in 2009, 2010, and 2011, before reaching a major inflection point in 2012, when nine more teams joined the analytics crowd, and the total number of analysts in the league doubled to 22. Today, our best estimate is that about 65 analytics personnel work in the NBA, and every single one of the 30 teams in the league employs at least one. The rise, when plotted on a graph, is literally exponential:

So the growth in NBA analytics professionals has been dramatic, but has it shown up where it matters, in wins and losses? In a word, yes. For this part of our analysis, we found it useful to divide the last eight years into two periods: 2008-12 (the left-most part of the graph above), and 2013-16, when things really took off.

Let's start at the beginning. Over the first phase of the analytics boom (2008-12), 10 teams employed zero full-time analysts, while the remaining 20 franchises employed at least one.

As it turns out, the 20 teams who managed to hire one or more full-time analytics staff by 2012 averaged about 7 more wins per season from 2008 to 2012 than the teams who hadn't. Sure, there are a lot of confounding variables at play here—some teams were going to suck, regardless, and perhaps the better-managed teams overall were also early adopters—but it's still an interesting result to report, even if we can't claim statistical significance.

More interesting, though, is this: the effect maintained itself. During the second period we considered (2013-16), those same 20 early adopters—the teams with at least one analytics professional by 2012—were still better off. Those vanguard teams averaged eight (!) more wins than their late-adopting opponents from 2013 to 2016, which is statistically significant and also, like, amazing.

The benefits showed up in specific tactics, too. Check out this chart, which shows the average number of three-point attempts per game from 2008-16.

Look familiar? It should. It's basically the exact same chart as we showed you a few paragraphs ago, about the rise of analysts, except lagged by about a year. Short story, then: As the number of analysts in the NBA has increased, so, too, has the number of three-pointers NBA teams are attempting per game. That makes sense, since pretty much all modern basketball research suggests that three-point attempts are a pretty good use of an offensive possession. As the NBA has entered the Age of the Analyst, it also has entered the Age of the Trey.

You have to execute, however, and our research shows that teams heavily staffed with analysts are doing that better than their opponents, too. Analytics-savvy teams had almost exactly the same three-point percentages as their conservative colleagues from 2008-2012: 35.7 percent for teams with an analyst, and 35.6 for teams without. From 2012-16, however, the early adopters shot threes 0.7 percent more accurately than the non-early adopters. That's a small difference, sure, but it is statistically significant, and it's not nothing. At the current rate of three-pointers attempted, that accuracy difference translates to a roughly 0.5 points per game advantage.

The same trend shows up in rebound differential, or how many more rebounds per game teams are getting than their opposition.From 2008-12, early adopters had a rebounding advantage over later teams, with an average rebound differential of +0.061 for the analytics-heavy teams and -0.12 for the non-analytics teams. That differential shrinks a bit later on: from 2012-16, the analytics teams got 0.075 more rebounds per game than the opposition, but the late adopters had an even rebound differential. For the early adopters from 2008-2012, that's about an extra possession every five games, or 15 extra over the course of the season. That can make a difference.

Perhaps the teams being left behind noticed the cumulative effect of all those threes and rebounds, because from 2013 to 2016, the number of analysts per team almost tripled. Which brings us to today's situation, wherein every single NBA team has at least one analytics staffer, and the average team employs just a little bit more than two.

But when everyone's doing something, it becomes relatively less advantageous to be doing it. During the second phase of our sample period, the returns on analysis dropped significantly: the 13 teams who had more than two in-house analysts actually won 1.9 fewer games per year than the 17 with two or fewer—although this was driven almost entirely by the performance of the 76ers (hey, Chuck, look at that!), who staffed up heavily during this period and proceeded to lose about 60 games per year in their effort at tanking. But that's no argument for falling behind the state of the art: the six teams with only a single analyst on staff during the second period fared eight wins worse per year than the rest of the league.

Now, there are all sorts of caveats here. The most important is this: the single biggest determinant of team success remains, more than any analytical capability or staffing structure, the quality of the players on the team. If your locker room is filled with 13 Elliot Williams's, you will probably suck no matter how savvy your staff is. And a team of 13 Charles Barkleys, safe to say, could probably beat most anyone pretty handily with nary a number-cruncher in sight. The difference we're talking about here is on the margins—improving teams over what they otherwise could have been—and that effect is mostly realized through changes in strategy driven by a deeper understanding of what plays and shots work more often than not, and by changes in personnel driven by a focus on one strategy over another.

Even taking that into account, though, our data strongly suggests a simple truth: that analytics work. In an 82-game season, the seven wins per year the early adopters picked up over their opponents every season from 2013-16 represent nearly nine percent of the total available. That's huge—the difference between a third seed and a seventh seed, and between playing into the summer and ending the year early. The old guard of the league may not like it, but change has come to the NBA in a big way, and if you're not part of it, you're generally being beaten by it.

(Editor's note: This piece has been updated to reflect more accurate data about NBA front office hirings).

Want to read more stories like this from VICE Sports? Subscribe to our daily newsletter.