Nylon Calculus: Investigating the rise and effectiveness of small ball

The NBA is a continuously evolving league. Rule changes, coaching innovations, and influxes of new talent have caused stylistic shifts in the way the game is played. Teams are shooting more 3s than ever, are trying to push the pace, and are opting for more skill and speed in their personnel on the floor. Gone, for now, are the days of low-post behemoths soaking up possessions. Ball-movement and spacing are the order of the day.

This shift has trickled down to lineups. As teams opt for more and more skill on the floor at once, they are naturally putting out smaller lineups. This is known as small ball, and its rise has been dramatic.

If we categorize players into offensive and defensive positions using clustering techniques on NBA.com’s four-year player tracking data we can visualize the rise of small-ball. Positional separation was by far the highest with three clusters, and this strongly corresponded to a ball-handlers/wings/bigs model on offense. I used a simple definition for distinguishing between traditional and small-ball lineups: If a lineup has two or more bigs, it is “traditional”. If there were one or fewer bigs, it is a “small-ball” lineup.

Teams have been going small in the playoffs more often since at least 2014, although this was still a clear minority of the total minutes played. We can see a significant increase in small ball beginning in 2016, especially in the playoffs. This trend continues into 2017, where over 37 percent of the minutes during the regular season featured small ball lineups, and for the first time, a majority of playoff minutes were played with small ball units.

Let’s zoom in on 2017 playoff teams:

The majority of teams went smaller in the playoffs — some dramatically so. Boston jettisoned Amir Johnson (who once again had a strong regular season by on/off metrics) from the starting lineup and his frontcourt pairing with Al Horford, who shifted up to play the center position exclusively. Oklahoma City, in an effort to give Russell Westbrook more spacing, largely abandoned their double-big lineups. Why did this happen? Maybe a lot of what “works” in the regular season might not in the playoffs. Or perhaps coaches believed that their best lineups were small ball units all along and felt the need to employ them more often in the playoffs.

Regardless, the rise of small ball should naturally be accompanied by evidence that small ball happens to beat traditional units when matched up head-to-head. To investigate this, I used the NBA.com player impact data, which allows for head-to-head comparison. After running all lineup combinations found in play-by-play data for the 2017 NBA season, I discovered that small ball units tended to outplay traditional lineups when matched up with one another.

The trend sharply reverses in the playoffs, although the sample size is many times smaller than that of the regular season. The combined score shows that small lineups bested traditional ones by about 1.2 points per 100 possessions in 2017.

Next: The NBA's super team era has placed championships over individual awards

Ultimately, the combination of NBA rules and talent dictates that the offensive benefit gained by going small generally outweighs the defensive cost. However, it is important to keep in mind that this is just a tendency in a single year and is largely dependent on team personnel. For now, however, the apparent effectiveness of small ball indicates that this trend will likely accelerate.

You can view all lineup combinations and their classifications here. The vast majority of classifications are sensible, but for what it’s worth, I imagine the two-year play-type data put to excellent use by Todd Whitehead would yield slightly more intuitive results.