Exploring the World of Assists, Passing and Ball-Movers

Title

If you asked a bunch of knowledgeable NBA fans about the best passers in the league, it wouldn’t be long before John Wall’s name came up. His assist numbers are eye-popping (with a career 9.2 assist per game average that’s good for 6th all-time) and he constantly creates easy shots for his teammates. Whether he’s whipping cross-court passes to open shooters or dishing to his pick-and-roll partner for an easy layup, Wall’s passing ability remains his greatest skill and the primary fuel for the Washington Wizards offense.

Consequently, things looked grim for Washington when it was announced in late January that Wall would miss six-to-eight weeks after a knee procedure. To everyone’s surprise, however, the Wizards have put together one of their best stretches of basketball this season, winning 12 of 19 games without Wall. Moreover, the WAY they’ve played without their star PG has been more fascinating than the impressive record. Wall may be one of the league’s best passers, but his team’s passing and assist numbers have spiked without him.

Wall has missed two chunks of time this year (once in December and his current prolonged absence), and the table below shows how the Wizards’ passing stats have fluctuated with and without him.Table 1 v2 - WAS StatsFirst, take note of the December stretch without Wall. The Wizards went 4-5 during that period and their offense sputtered to the tune of 101 points per game on a 49% effective field-goal percentage (numbers you’d expect from one of the league’s worst offenses). They passed the ball much less, and the team is not built to run smoothly that way without Wall running the show.

During Wall’s most recent absence, however, they’ve played a different brand of basketball and have been an excellent offensive team. Over that period, Washington has stylistically looked more like the Golden State Warriors than the Wizards we’re used to seeing. John Wall’s strengths and weaknesses have a lot to do with those differing styles, but we shouldn’t fall into the trap of misinterpreting the aesthetic differences for improved results.

Wall is a hyperathletic point guard with terrific court vision who struggles without the ball. As a result, much of Washington’s offense with Wall looks like this:

Pristine spacing, a simple high screen-and-roll, blazing speed and no easy choices for the defense. The Wizards don’t need eight passes to find a good look here; one pass gets them a dunk or a three.

Without Wall, Washington has had to improvise much more. Bradley Beal and Tomas Satoransky have been phenomenal, but they don’t provide Wall-levels of dynamism with the ball in their hands. When initial actions are stymied, the team falls back on ball movement, player movement and swinging the ball around until they create an open look:

When we think about great passers, we normally think of players who resemble the John Wall archetype: ball-dominant players who probe, read the defense and actively create shots for teammates (e.g. Chris Paul, LeBron James, Steve Nash). The irony of this type of passer is that they can often limit the ball movement on their teams. The Wizards’ recent run of success is a good example of a different kind of passing centered on quick decision-making and a high pass frequency. This type of decisive passing is incredibly important for teams that don’t possess elite creators like Wall.

The traditional assist statistic isn’t perfect1, but it’s pretty good at highlighting the Wall archetype of passer. The main problem with the assist is that it doesn’t reward the ball-movers—the players who add offensive value by making quick decisions and throwing the simple passes that keep defenses on their toes. Fortunately, NBA.com publicizes some of the tracking data provided by Second Spectrum, so I explored these statistics to learn more about the wonderful world of passing and ball movement.

[Note: This analysis uses “Assists Adjusted” (noted as AST ADJ on NBA.com) instead of “Assists”. Assists Adjusted include assists, secondary assists (the hockey assist) and free-throw assists (passes leading to shooting fouls).]


Passing & Ball Movement on a Team Level

Spurs title

The 2014 San Antonio Spurs are very high on my list of favorite playoff runs. It didn’t matter that they were old and lacked a superstar (this was before Kawhi Leonard emerged as the NBA’s best cyborg). They were a buzzsaw of passing and cutting that demolished the opposition by embodying the spirit of the jogo bonito (the beautiful game). The next season saw the rise of the juggernaut that is the Golden State Warriors, who play with a similar type of joy derived from plentiful passing and an aversion to isolation basketball.

Despite the recent success of these two teams, passing and ball movement—beautiful as they may be—don’t strongly correlate with improved offensive performance on a league-wide level. The tracking data related to passing only dates back to 2013-14, but since that season there is moderate positive correlation between assists and offensive rating. Beyond that, factors like passes, potential assists, seconds per touch and the percentage of field goals that were assisted only show a weak relationship with offensive rating.Table 2 - League-wide CorrelationThis year’s Houston Rockets are currently posting the highest offensive rating in history (115.8 points per 100 possessions), and they’re second-last in the league in passes per game while setting records for isolation frequency and efficiency. Their offense is brutally simple: let James Harden and Chris Paul go to work.

As with most things in the NBA, context is everything. The Rockets and Warriors are the two best offenses in the league, but they utilize different approaches because of their stars. Harden and Paul are both excellent shooters, but they are most dangerous with the ball in their hands running pick-and-roll or operating in isolation. Conversely, Stephen Curry and Kevin Durant can make plays off-the-dribble, but they are most threatening in a motion-heavy Warriors system that leverages the threat of their shooting (along with fellow Splash Brother Klay Thompson).

As I move on to analyze individual player statistics, keep this mantra (context is everything) in mind. Whether the statistical profiles reveal a deliberate, ball-dominant passer like Rajon Rondo or a decisive, hot potato ball-mover like Draymond Green, one archetype isn’t necessarily better or more valuable than the other. It all depends on their complementary skills and the roster/system around them.


Individual Assists & Passing

NOTE: Please read Footnote #2 for more information about how I obtained and filtered this data.

Assist Volume and Efficiency

Graph 1 - Adj Ast to Pass % vs. Adj Ast per game

The top-right quadrant of this graph is where the traditional “best passers” come out to play3. These are the assist kings, and most of them are such good passers that their team’s offense is built around their ability to create shots for teammates. They typically haven’t needed a ball movement system around them; just give them the rock and let them do their thing. As a result, they land very high on the adjusted assist-to-pass percentage axis as well.

Assist-to-pass percentage refers to the percentage of the player’s passes that are assists. Russell Westbrook leads the pack with a whopping 20 percent of his passes resulting in assists, and the top of the list is littered with names you’d expect.

Adj Ast-to-Pass % TableA few surprising names: Will Bynum, Kobe Bryant and Kevin Durant. Bynum last played in 2014-15 and barely qualified for my playing time threshold (1,121 mins played), but his per-minute numbers are surprisingly good. He used his speed, handle and power to attack the rim and he had pretty good passing vision when he was able to collapse the defense. It was the other aspects of the game (namely shooting/finishing) that held Bynum back, but these numbers suggest he was a better passer than I originally thought.

It’s important to look at assist-to-pass percentage in conjunction with assist volume (Adjusted Assists per game, in this case) to provide more context. Kobe Bryant has the 13th-highest adjusted assist-to-pass percentage in the sample, but he only racked up 4.7 adjusted assists per game over this period (63rd of the 148 players).

That’s actually more adjusted assists per game than Manu Ginobili, but Bryant did it in about 10 more minutes per game—a factor that is revealed when we compare their numbers in the per-minute categories explored later in this article (see below). Ginobili is 8th in assists per minute of possession and 26th in passes made per minute of possession; Bryant is 66th and 141st respectively. That assortment of metrics suggests that Bryant dominated the ball, was able to make productive passes (i.e. ones leading to assists), but he didn’t do this at a high rate.

Then we have Kevin Durant. Looking at the range of statistics analyzed here, I think I’ve underestimated how good Durant is as a passer, creator and ball-mover.Kevin Durant Passing sThis data captures his performance with the Thunder and the Warriors, so it’s not drastically skewed by playing in Golden State. There are only two players in the sample who are in the top 40 for each of the above categories: Nicolas Batum and Kevin Durant (we’ll get to Batum in a second). Durant’s ridiculous scoring ability (and the speed with which he gets buckets) is the hallmark of his game, but he’s able to punish defenses with his passing if they commit too many resources to stop his scoring.

There aren’t too many shockers in the top- and bottom- 15 in adjusted assists per game, but the table highlights the potential impact of team scheme on these numbers.Adj Ast per Game TableBen Simmons and Joel Embiid are represented on either end of the spectrum here, but one thing worth noticing is their adjusted assist-to-pass percentage relative to the rest of the players in the table—both are lower than their peers. One potential factor contributing to this phenomenon is how much the Philadelphia 76ers pass the ball. The Sixers pass the ball more than any other team in the league, and their offense is built around constant ball movement. There is so much passing built into their sets that it may have dragged down their numbers here.

The Ball-Movers

Graph 2 - Passes per min vs. Ast per min

The top-right quadrant of this graph highlights some of the best ball-movers in the NBA. These players pass at a high rate, don’t hold on to the ball for too long, and create points with their passing vision. All of them generate significant offensive value for their teams because they make quick decisions with the ball and set up teammates.

The fact that most of those players are big men reveals some of the underlying biases of these tracking stats. Since traditional point guards spend so much time with the ball, they suffer in these per-minute-of-possession categories4. On the flip side, big men typically don’t have the ball for that long and can rack up passes because of their rebounding (after a rebound, they pass to a ball-handler to bring it up the floor)5.

These confounding factors make Nicolas Batum’s numbers even more impressive. Not only is he a quality ball-mover (as shown in this graph), but his 6.4 adjusted assists per game is 27th in the sample, ahead of players like Tony Parker, T.J. McConnell and Blake Griffin. By these metrics, Batum is one of the special passers in the game, capable of racking up assists without soaking up seconds of possession.

Here are the top- and bottom-15 for the axes of the above graph:Adj Ast Per Min

Passes per MinThere is a lot of overlap between the top 15s of each table, and the numbers align with the eye-test here. Draymond Green, Nicolas Batum, Joakim Noah, Boris Diaw, Al Horford, Pau Gasol, Marc Gasol, Nikola Jokic, Joe Ingles, Tim Duncan and Josh McRoberts. All of those players pop on film as good decision-makers who keep the ball moving.

There is less overlap between the bottom 15s. The bottom 15 in adjusted assists-per-minute contains a number of score-first primary ball-handlers, so their per-minute numbers likely take a hit because they spend so much time looking to score with the ball rather than creating for their teammates.

The bottom 15 in passes made per minute reflects the ball-dominant players—the ones where the ball sticks in their hands with regularity. As I mentioned before, however, that’s not necessarily a negative. Russell Westbrook, James Harden, Damian Lillard and Chris Paul are all excellent primary ball-handlers, so you want the ball in their hands as much as possible.

 

There is still much more research needed on this topic, but the NBA.com tracking data is a step in the right direction towards understanding the different types of passing and the value each one can generate. Further areas for exploration include analyzing team-level passing stats when various players are on/off the court, and perhaps including some of the speed/distance data made public by the NBA. Ball movement is so closely intertwined with player movement, that accounting for players who move more on the offensive end might shed more light on the subject.


1. I recommend reading Ben Falk’s musings on the shortcomings of the assist (subscription required).^

2. Note about data

All data was obtained from NBA.com/Stats. This tracking data goes back to the 2013-14 season, so I compiled all individual statistics for every player in the league from the 2013-14 season till the 2017-18 season (with the All-Star Break, 2/16, as a cut-off for this year). I then filtered the data to only include players who met the following criteria:

Minutes played ≥ 1,000

Minutes per game ≥ 15

Adjusted Assists per game ≥ 3

This left me with my sample population of 148 players.^

3. Russell Westbrook is probably underrated in this regard. He doesn’t have the same level of creativity as Chris Paul, John Wall, James Harden and LeBron James, but he is just as productive as a passer because he leverages his scoring and athleticism to devastating effect.^

4. The simple act of bringing the ball up the floor adds up over a season and inflates their time of possession numbers.^

5. One of the reasons I implemented a minimum of 3 AST ADJ/game for my data pool is that there was too much noise with the per-minute passing stats skewed heavily towards big men.^

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s