Some Fancy Passing Stats
By Seth Partnow
![Mar 20, 2015; Los Angeles, CA, USA; Los Angeles Clippers guard Chris Paul (left) drives the ball defended by Washington Wizards guard John Wall (right) during the fourth quarter at Staples Center. The Los Angeles Clippers won 113-99. Mandatory Credit: Kelvin Kuo-USA TODAY Sports Mar 20, 2015; Los Angeles, CA, USA; Los Angeles Clippers guard Chris Paul (left) drives the ball defended by Washington Wizards guard John Wall (right) during the fourth quarter at Staples Center. The Los Angeles Clippers won 113-99. Mandatory Credit: Kelvin Kuo-USA TODAY Sports](https://images2.minutemediacdn.com/image/upload/c_fill,w_720,ar_16:9,f_auto,q_auto,g_auto/shape/cover/sport/5b00592a1d7ca1471701a863601644e2b46b1fe37b7b5bd3df046b703543d3f8.jpg)
Over at WaPo’s #FancyStats, I took another look at a question to which I’ve returned to over and over again which is measuring playmaking beyond simply counting assists. I’ve felt like the “Assist Usage” component of True Usage does a decent job of measuring activity, but measuring achievement in terms of decoupling the skills of the passer and shooter have felt pretty intractable for a while. A big stumbling block has been the lack of passing/assist data on missed shots. I’ve bemoaned this a little before, but that one simple addition to SportVU or PBP logs would make all sorts of comparisons possible for both playmaking and individual shot creation.
In any event, there wasn’t room in the Post article for any sort of data-dump, so I figured I’d drop the overall numbers here. Below are the estimated Playmaking TO%, Playmaking Usage and Playmaking eFG% for every player in the league who played more than 2000 possessions with any one team. I’ve found that changing team contexts has a pretty big effect on these types of stats so I’ve purposefully not included season totals for players who switched teams mid season. A few miss the 2k possessions cutoff (Dragic in Miami, Rondo in Boston, etc.)
While these are all estimates rather than exact values given the imperfections of the underlying passing stats, Playmaking TO% in particular should be taken with a grain of salt as “Bad Pass” turnovers both under and overcounts what might actually be turnovers committed in the course of playmaking. Some passes which are thrown away weren’t going to generate assists on one hand, while on the other, there are plenty of lost ball and offensive foul turnovers which do arise from situations where a player is trying to set up a teammate. Still, it’s a closer approximation than raw turnover totals, so I’ll live with the imperfections until better data comes along.
[Quick Update 7/28/15]: Based on a twitter convo, decided to provide a little context for the eFG’s listed below, especially because big men somewhat appear to be “better” passers. This is large part, I think, due to the tendencies of where the passes are going as much as the players making passes. In general terms, point guards are doing things like hitting big men for mid range jumpers in pick-and-pop situations while big men aren’t doing so. In effect, this means bigs are getting “easier” assist looks much like low shooting usage players tend to be shooting more open, higher general efficiency shots than stars. The chart below illustrates what I mean based on the partial season data I first used here and here:
It’s not a huge difference, but those assists to the midrange “Donut of Doom” are on balance far less efficiency than “MoreyBall” friendly rim attempts and three pointers.
Here’s the full chart: