With the NBA’s pendulum swinging swiftly from an emphasis on isolation scoring toward a more collective mode of shot creation, making plays for teammates has become increasingly important. It is well-established that scoring attempts resulting from the collaboration between teammates are much more efficient than individual play.
But that’s just the result. How that cooperation comes about, and more importantly, who is best at making it work are questions that haven’t been especially well examined. Assists tell us somethings, but they can be irregularly awarded and often say as much about a player’s teammates as his own passing ability. As an example, the relationship between usage and efficiency in shooting has been studied almost to death.
However, as the ability to track individual contributions at a finer level of detail improves, those sorts items can at least be addressed. It is now possible to measure what I’ve termed “Assist Usage,” or in other words: how often does a player set up a teammate?
In 2014-15 these were the top 10 players (minimum 3000 possessions with approximately 1,500 minutes played with a given team):
Not a lot of terribly surprising names. But that is only a measure of activity, not achievement. Much like Westbrook himself puts up a ton of shots, merely having the ball enough to create those shooting opportunities doesn’t say much about how well a player performs in those creation situations. Though as assisted shots strongly tend to be better than unassisted ones, the quantity of assists chances has a quality all its own.
Still, knowing the most active playmakers isn’t quite the same thing as identifying the best. There has to be a measure of efficiency or effectiveness to go with volume. One possibility is to simply take the effective field goal percentage on a player’s assist chances, (among players with at least 200 assist chances for a team last season):
A few things to note. First is the presence of a lot of big men, and more specifically, the presence of a lot of guys with good three point shooting around them. Aside from the actual names and numbers on the list, the above could still be as much about teammates making shots as anything else. Further, to only measure “successful” assist chances, where the intended shooter actually received the ball and did so in position to get a reasonable shot off, is missing some important information. After all, the worst thing that can happen when trying to set a teammate up is not that they miss the shot, it’s that the pass never gets there.
Measuring “playmaking” turnovers is difficult, as that information is not explicitly tracked anywhere in the public domain. However, there is a shortcut. Various play-by-play accounts include whether a given turnover was the result of a bad pass, a lost ball, an offensive foul as well as other possibilities.
It’s not perfect, but if “bad pass” turnovers are assumed to be “missed” attempts at playmaking, how does that change things?
We can estimate which players are some of least and most likely to turn the ball over trying to create for teammates (NBA average was 12.7 percent in 2014-15):
Refiguring the passing efficiency stat while including those turnovers probably gives a better estimate of passing effectiveness than only those passes which found the target — we would never attempt to measure a player’s ability as a scorer by only counting the times he was able to get himself wide open.
This now allows for a rough analogy of the scoring “usage-efficiency” trade off. Note the “usage” below includes bad pass turnovers to allow for a fair comparison with the turnover-including efficiency numbers (min 400 assist chances):
For comparison’s sake, the NBA averages for “Playmaking Usage” and “Playmaking Efficiency” under this method were 10.4 percent and 48.9 percent respectively.
So who is the best playmaker? It’s hard to suggest anyone other than Chris Paul given his high rankings in both usage and efficiency, not to mention his absurdly low turnover rate. Holiday’s steadiness and the scoring dominance of James and Curry also give them top marks, while Blake Griffin’s own exceptional vision and ball-movement skills are illuminated as well.
On the other hand, this exercise demonstrates part of why Josh Smith can be such a maddening player: his passing can be a total adventure as evidenced by his astronomical turnover rate in Houston. At the same time, whether in Houston or Detroit, when he finds teammates, they get good shots as shown by the raw efficiency off his passes. In part it is because he seems to be excellent at finding teammates for high value layups and threes. During his time in Houston, nearly 94 percent of his assists (and 88.6 percent overall including his time in Detroit) went to players shooting either at the rim or beyond the arc. It is this ability which makes him such an intriguing addition to the Clippers’ bench this coming season.
As a final note, the shot making ability of teammates hasn’t been accounted for at all. For a number of reasons this is an exceptionally difficult problem both from a logical and data standpoint. Who is properly credited when a great passer delivers a perfect set up to a great shooter in rhythm? How good a pass is it to find a player the defense willingly leaves open? How can the quality of a pass be measured given the tools at our disposal. These are questions for future steady (not to mention fuller data), but for now, this is a start.