Playmaking and Untangling Individual Contributions From Cooperative Activity

Mar 1, 2015; Chicago, IL, USA; Los Angeles Clippers guard Chris Paul (3) passes with Chicago Bulls forward Pau Gasol (16) defending during the second half at the United Center. The Clippers won 96-86. Mandatory Credit: Dennis Wierzbicki-USA TODAY Sports

Earlier this week, I took a first look into which shots are assisted and from where on the court those passes originate. In examining league-wide data, some trends were identifiable in terms of the importance of assisted shots and the desirability of running offensive from the middle of the floor and not the corners. In this installment, I wanted to look at some more individualized points.

Before I get to that, an overarching caveat for nearly everything that follows — values are generally going to be assigned to the passer. Al Horford’s assist chances have had the highest effiiciency of those of any player in the league. Zach LaVine has thrown a higher proportion of his assist chances to midrange areas than any other player, while players have shot horrifically from three on dishes from LaMarcus Aldridge. Those plays are identified with the passses as much out of convenience as from any real attempt to designate skill.

In Horford’s case, playing in the Hawks’ offense with the Hawks’ shooters certainly has something to do with his apparent passing efficiency. By the same token, he’s also a good passer by reputation and observation and the high ranking does nothing to contradict. On the other hand, LaVine has spent a lot of time playing with a banged up Wolves roster in an offense somewhere between poorly-designed or expertly designed to get the best bad shots possible while Aldridge might simply be the victim of bad luck in terms of guys just missing shots. With the data available, parsing between these contextual and random factors and pulling out a player’s skill element in any sort of systematic way is probably not workable as yet. But we still can learn something about how a player operates at least within his current team concept and roll.

With that out of the way, here are some of the more intriguing playmaking findings:

1. Getting into the paint appears to be a good thing: In Wednesday’s post, we looked at how effective assists from various portions of the court can be. Remember, these represent assists marked by where the pass originated from, not where the shot was taken:

Playmaking from the lane area, whether inside the restricted area or not, tends to lead to accurate shots. Unfortunately, the data[1. Courtesy of the magical SportVU elves.] is not fine enough to see which shots are being assisted from which areas. However, by combining this information with a version of my XeFG[2. Expected Effective Field Goal Percentage.] which measures the shot quality of a each player’s assists, there does appear to be a strong correlation between players who assist relatively more from the paint and those who’s assists find the most efficient shots:

There’s clearly a lot of noise in there, and the correlation should not be mistake for causation[3. The space needed for the player to get into the lane might also necessarily provide shots being more open off the resulting pass as both are effects rather than causes of good, efficiency offensive sets and movement.], but getting into the lane does seem to overlap with more open, higher value resulting shots. Not a revelation to existing basketball theory, but it’s nice to see the numbers and the theoretical wisdom overlap some. Similar analysis shows the percentage of assist chances coming from inside the arc but outside the lane not really affecting the openness of assisted shots either way, while a higher propensity to pass from outside the arc being slightly negatively correlated with the expected efficiency of resulting shots. Even that analysis is colored by the difference in rolls — most of the players with high “outside” assist rates are jump shooters and point guards, while those assisting relatively more from in close tend to be big men. One notable exception is Tony Parker, the only qualifying point guard with fewer than 20% of his assist chances having been generated outside the arc.

Some questions are simply impossible to address with only this year’s data. Kevin Love’s transformation into the billionaire’s Channing Frye has been accompanied by most of his playmaking occurring from the wings as opposed to the time he spent facilitating from the low post or elbow last season in Minnesota:

Minimum 175 total assist chances through 3/3/15

Or at least that’s the theory, without similar data from last year, it’s hard to compare. Though that could be a very useful impactful use of this sort of analysis — Derrick Rose is in the top quartile of proportion of assist chances from outside the arc this season, which was probably not close to the case during his pre-injury heyday. This would be another way to identify the change in style and perhaps could be an indicator of a return to form if a player’s “paint assist chances” started to rise again.

2. Team concept shows. The “MoreyBall” philosophy of layups, threes and free throws is great in theory, but as these charts from Wednesday indicate, getting those threes and layups consistently take input of multiple teammates:

Three pointers are heavily reliant on being set up, whereas there is a massive efficiency gap between assisted and unassisted shots at the rim[4. Some of this gap probably represents breakaways and fast breaks, but the gap is too large for that to be the only explanation.]. So it follows teams wishing to maximize shots from those locations need to pass to ball to teammates shooting from those locations. Looking at individual “assist to” locations, the top of the leaderboard of layup-and-three friendly assists has a certain sameness to it:

Minimum 100 potential assists through 3/3/15 * Includes time with Detroit ** Includes time with Phoenix

Probably not surprising, but still, illustrative of the Houston plan of attack in action[5. League average is around 67% of assisted shots attempted are from restricted area or three point range]. The other end? Pacers and Timberwolves[3. Sorry Zach LaVine] and Wizards, oh my.

3. A stat which I’m not really sure means anything. Among the top 200 in assist-chancers here are the top and bottom 15 in eFG% on those potentially assisted shots:

One interesting thing to note playing around with this and a bunch of other passing data is almost anyway it’s massaged or sliced, Josh Smith comes out looking really good, even with a sizable chunk of his minutes being from his time in Detroit. Other than that, what to take away from this? It’s tempting to see names like Horford, Nowitzki, Griffin, and James and decide some playmaking code has been cracked, but I’m not sure I’d place that much weight on it. Aldridge is an illustrative counterexample. Prior to his unfortunate Achilles injury, Wes Matthews was shooting 38.8% on threes for the year. Yet in the timeframe this data is current through[5. Games of March 3, 2015], Wes was only 12 of 48 on passes threes after receiving a pass from Aldridge. Which is more likely, that Aldridge is in the habit of giving especially poor passes, or Matthews has simply had a rotten run of shooting when he happens to receive a kick out from Aldridge? Could it be some of both? Since we (still) know very little about potentially assisted but missed shots it’s hard to say.

Playmaking is a bit like shot defense in this regard — the interaction is so complex as to why a given shot goes in or doesn’t we should be somewhat reticent to make sweeping claims about individual skill when an invariably small, likely biased sample of shots seems to indicate something about individual passing skill. Much like good coaching usually involves having good players at your disposal, the best way to be a good passer is to play with good shooters.