# Nylon Calculus: The Rockets’ consistent 3-point threat

Even with increased 3-point shooting leaguewide, the Houston Rockets have continued to stand out for their frequent attacks behind the arc. For the second consecutive season, they have taken roughly half of their field goal attempts from downtown, outpacing their closest competitor (the Milwaukee Bucks) by about 10 percentage points and exceeding the NBA average by over 15 percentage points. Yet it isn’t merely a matter of volume. Seldom discussed is how consistently the Rockets have attempted 3-pointers across different offensive contexts.

It might seem like a given, but it’s a more nuanced subject than we might initially assume. After all, different offensive contexts beget different types of scoring opportunities, and teams may tailor their strategies accordingly. Consider two basic scenarios: halfcourt offense and transition offense. With defenders generally more organized in the former, high-value shots (particularly those at the rim) tend to be more abundant in the latter. Just think of how much easier it is, in relative terms, to get a dunk on a fastbreak. It helps illustrate the idea that typical shot profiles vary by circumstance. Indeed, one of the crucial limitations of standard shot charts — apart from overlooking free throws — is their aggregation of multiple offensive contexts that present fundamentally different looks at the basket. Not all layups are created equally.

Of course, for a fluid and dynamic game like basketball, discrete offensive contexts are hard to define. Ideally, we’d use tracking data to identify specific states, like early offense, and generate shot statistics from there, but these details are not publicly available. Synergy leverages video analysis to diagnose common play types such as “Pick-and-Roll – Ball Handler,” “Pick-and-Roll – Roll Man,” “Dribble Handoff,” “Spot-up” and others. While summary statistics have previously been posted online, they are currently off NBA.com. Cleaning the Glass has a “Play Context” table disaggregating general efficiency data by “Halfcourt,” “Transition,” “Putback” and “Miscellaneous” plays, though shot zones are not among the variables. Even when definitions are achieved, data limitations make them difficult to incorporate into broader research.

One simple solution is to parse play-by-play logs and categorize true shot attempts (i.e., field goal attempts and trips to the free throw line, excluding technicals) by how a possession starts. Put differently, what’s the preceding event at the opponents’ side of the court that sets the stage for the scoring opportunity? It’s an approach that Inpredictable employs to calculate per-possession statistics and that pbpstats.com features in its modules, providing a glimpse into various conditions that may shape a team’s offensive options.

In the following analysis, I break down true shot attempts into four areas: those that follow made shots, missed shots, dead balls and live ball turnovers. Dead balls include turnovers, violations and team rebounds that require the ball to be inbounded as the possession switches. Live ball turnovers involve steals. Collectively, these categories reflect the main defensive outcomes from which teams subsequently generate their own offense.

There are shortcomings, to be sure. For one thing, made shots lump together both converted field goals and converted free throws, even though a reasonable argument can be made that the set positioning of players during the latter creates a different situation altogether. For another, a missed layup and a missed 3-pointer may lead to varying dynamics as the defense rebounds the ball and takes it the other way. Furthermore, putbacks, timeouts and other events within a possession, which can substantially alter the complexion of a scoring opportunity, are not teased out. The four categories certainly have their advantages in terms of simplicity, but they admittedly overlook a number of valuable nuances.

Nonetheless, they serve as a decent springboard for shooting breakdowns. The following heatmaps, for instance, display the locations of all true shot attempts in the NBA this season based on the start of the possession. Note, in particular, the restricted area and the 3-point zone.

Shot frequencies at the rim and behind the arc are relatively stable across contexts, except for live ball turnovers. In this state, there’s increased activity in the restricted area and reductions elsewhere, which makes intuitive sense when we consider the types of fastbreak opportunities that become available after steals.

As with other teams, the Rockets’ 3-point frequency declines. Whereas they’re closer to 50 percent after made shots, missed shots and dead balls, they drop to 41 percent after live ball turnovers.

On the one hand, we can certainly see the Rockets shifting down toward the pack within the context of live ball turnovers. On the other hand, even with the slide, they still locate two of every five true shot attempts from downtown — a higher rate than what nearly every team puts together in the halfcourt and other situations that are typically more conducive to 3-point shooting. Only Milwaukee’s frequency after missed shots is on par.

Speaking of the Bucks, they serve as an illustrative contrast to Houston’s consistent 3-point attack. They have a top-three rate after made shots, missed shots and dead balls, but dip below average after live ball turnovers. That’s because they ramp up scoring opportunities at the rim, where their 64 percent frequency is surpassed only by the Los Angeles Lakers’ 65 percent. When you have Giannis Antetokounmpo, who operates exceptionally in transition, it’s a great strategy.

The Rockets’ comparative advantage is different. With two masterful halfcourt playmakers in James Harden and Chris Paul, they can generate efficient looks without necessarily pushing the ball up the court. Their offensive pace after steals is indicative of this idea. According to Inpredictable, their 9.5-second average places them among the five slowest teams for the second year in a row. They’ve actually hovered fairly consistently around the 9- to 10-second range over the past five seasons, but their relative standing has slipped as the league has trended toward a faster pace after live ball turnovers. It would require a more detailed study, but my hypothesis is that there’s less urgency for a team like Houston to follow the crowd when its established approach and personnel already yield good shots, including 3-pointers.

In the True Shooting Charts app at positiveresidual.com, we get a glimpse of how the Rockets have performed by start of possession. Here’s a summary:

After live ball turnovers, the Rockets are at 67 percent true shooting, a top-10 clip. Approximately 11 percent of their attempts are corner 3s, where their true shooting is at 77 percent. The samples are small, so they are likely to revert toward the mean, but they nonetheless help lend some extra credibility to Houston’s consistent threat from deep.

*All stats current through 1/31