Nylon Calculus: Can a defense influence ball movement?

OAKLAND, CA - MARCH 8: Nikola Jokic #15 of the Denver Nuggets handles the ball against Kevin Durant #35 of the Golden State Warriors on March 8, 2019 at ORACLE Arena in Oakland, California. NOTE TO USER: User expressly acknowledges and agrees that, by downloading and or using this photograph, user is consenting to the terms and conditions of Getty Images License Agreement. Mandatory Copyright Notice: Copyright 2019 NBAE (Photo by Noah Graham/NBAE via Getty Images)
OAKLAND, CA - MARCH 8: Nikola Jokic #15 of the Denver Nuggets handles the ball against Kevin Durant #35 of the Golden State Warriors on March 8, 2019 at ORACLE Arena in Oakland, California. NOTE TO USER: User expressly acknowledges and agrees that, by downloading and or using this photograph, user is consenting to the terms and conditions of Getty Images License Agreement. Mandatory Copyright Notice: Copyright 2019 NBAE (Photo by Noah Graham/NBAE via Getty Images) /
facebooktwitterreddit

The Denver Nuggets have risen to the upper echelon of the Western Conference on the strength of an elite offense that features crisp ball movement. They assist on roughly two-thirds of their made field goals, trailing only the Golden State Warriors for the league lead. Their four secondary assists per game place them among the top five teams, and their 47 potential assists per game rank among the top ten. They convert passes to assists with notable efficiency. As Ian Levy showed in his offensive style charts, they move the ball quickly from player to player, averaging under three seconds per touch. They’re able to execute at this level not just because their franchise player, Nikola Jokic, is one of the premier facilitators in the NBA, but also because they have other key contributors, like Mason Plumlee and Paul Millsap, who are productive passers for their positions.

It is, therefore, no surprise that, when the Nuggets hit a rough patch recently, their waning ball movement was singled out as the primary culprit. According to a report on the team’s web site, head coach Michael Malone conducted a film session in which he highlighted undesirable shots that were preceded by fewer than two passes. Beat writer Harrison Wind broke it down even further at BSN Denver. He pointed out that, after averaging 312 passes per game before the All-Star Break, the Nuggets slid below the 300-mark. We can see this trend in their ten-game rolling average chart:

Note that there appears to be missing data in their March 6 game, which shows the Nuggets and their opponent, the Los Angeles Lakers, at implausibly low totals of 92 and 48 passes, respectively (that explains the very abrupt dip at the 60-game mark above). But, even if we assume that the teams had average passing performances that night, Denver would still be under 300 passes per game over their last 10 contests — their lowest point of the season.

Malone is far from the only coach to decry poor ball movement. Steve Kerr, for example, has been known to grumble whenever the Warriors fall short of 300 to 325 passes per game for prolonged stretches, and it’s common for teams to associate their failure or unwillingness to share the rock with offensive woes. Such proclamations hint at an implied assumption that, when it comes to passing, the offense largely controls its own destiny. If the ball isn’t popping, conventional wisdom seems to suggest it’s because the offense is departing from “the basics”; seldom is it attributed to defensive execution.

Yet I think it’s fair to wonder how much influence, if any, the defense has on ball movement. After all, the Nuggets’ stagnant passing has occurred during a period when they faced the Indiana Pacers, the Utah Jazz and the Oklahoma City Thunder — three formidable defensive units. They also had games against the Warriors and the San Antonio Spurs, whose defenses have been less imposing than their pedigrees would suggest, but can nevertheless put together vintage performances on a given night. It’s possible that, even with every intention of swinging the ball around the court, the offense may find itself deterred by competent opponents.

Intuitively, we can also conceive of on-court scenarios when the defense might dictate ball movement. Think of a play that calls for an entry pass to the post, but the defender fronts the intended receiver and compels the ball handler either to shoot or drive instead. If you have enough of these situations in which advantages are denied or turnover risk increases, you might see passing totals drop. Conversely, the offense might be able to initiate ball movement, but defenders rotate, recover and close out so well that they prevent shots from being taken and induce extraneous passes late into the possession. In such cases, good defense might actually raise the number of passes. Beyond the half court, consistently poor floor balance might lead to additional transition opportunities, which require limited ball movement to exploit, so bad defense could suppress passes, as well. The bottom line is that the data can shift in either direction based on factors outside the offense’s immediate control.

But how much sway does the defense generally have on ball movement? To answer this question, I draw upon analytics literature that examines how each side of the ball helps explain the variance in a metric of choice. In 2015, Ken Pomeroy explored offensive and defensive influences on commonly used statistics such as free throw percentage, average possession length and block rate in men’s college basketball. Inpredictable conducted a similar study on pace in the NBA. A few years earlier, Chase Stuart and Neil Paine had applied the concept to scoring in the NFL. Every one of these articles focused on the relative importance of offense and defense.

Accordingly, I use a simple linear regression model that attempts to predict a team’s passes for a particular game based on its season-long passing average and its opponent’s season-long defensive passing average. The variables are expressed as passes per possession and scaled to league average. The season-long numbers exclude data from the game in question. Overall, there are 14,330 observations from regular-season games between 2013-14, when tracking data were initially released, and 2018-19 (through March 16).*

The model explains 46.1 percent of the variance in passes per possession. Using the relaimpo package in R, we can then estimate the relative contributions of the offense and the defense to this result.

As the table shows, within the proportion of the variance in passes per possession that’s explained by the model, approximately 90 percent is attributable to the offense. If we assume that Pomeroy’s findings apply to the NBA, then ball movement is predominantly on the offensive end of the offense-defense spectrum — right around where average possession length lies. Defense certainly has its place, and the various scenarios enumerated above are potential manifestations of its influence. But, by and large, the offense is in the driver’s seat.

Admittedly, this analysis represents a preliminary attempt to make sense of ball movement. A better measurement than just passes per possession, which includes inbound passes after made shots and other activities that have limited relevance to core offensive strategies, is likely to improve our understanding. The same goes for a more sophisticated model. In general, though, it appears warranted for a team to look initially in the mirror whenever ball movement stagnates.

* Note that there are 26 games without passing data, or 0.4 percent of the total eligible pool of games.