NBA Offenses Got More Efficient by Giving Fewer Shots to the Most Efficient Players. (I’ll Explain.)
It may sound paradoxical, but the league has gotten better over time at assigning players their roles.
On a per-possession basis, there isn’t a more efficient option in the NBA this season, or maybe ever, than having Dallas Mavericks center Dwight Powell do something with the basketball. Whenever Powell finishes a play (by either shooting, getting to the line or turning the ball over), the Mavericks are averaging 151 points per 100 possessions. That’s not only 30% better than the league average — which is itself skyrocketing to new heights all the time — but it is also the best individual offensive rating by a qualified player in the history of the league.
And yet, with a usage rate of just 9.3%, Powell is also the second-most infrequently used offensive option in the NBA, ahead of only Nic Batum of the Philadelphia 76ers. Not coincidentally, Batum is also wildly efficient with his possessions, with an individual offensive rating of 139 points per 100 possessions. In fact, the bottom of the usage leaderboard is littered with players whose offensive ratings reach 120 or higher: Pat Connaughton (123 ORtg/10.2% usage); Dean Wade (123/10.4%); Kevon Looney (138/10.9%); Isaiah Hartenstein (134/11.1%); Al Horford (132/11.3%); Amir Coffey (136/11.4%); Cason Wallace (127/12.0%)… you get the idea.
None of this is coincidental. It’s actually part of a long-running trend whereby the correlation between who gets to use possessions, and how efficient they are with those possessions, has been steadily decreasing over time.
Over that same span, leaguewide efficiency has absolutely taken off, particularly in the past decade or so — a time when individual usage and efficiency decoupled more rapidly than ever.
How does this make any sense? How can the league become far more efficient while simultaneously allocating a smaller share of possessions to its most efficient players?
The way to square those trends is to think about efficiency as a function of usage — or what
long ago called “skill curves”. Powell is hyper-efficient in his small share of possessions precisely because he has a small share of possessions: almost all of his play types are either as a roll man, cutter, offensive rebounder (for putbacks), or getting fouled and going to the line. While Powell is great in that role, those are about as efficient a collection of play types as a player can have. Asking him to fill an expanded role would mean adding extra plays that require a wider range of skills and are not as high-efficiency at baseline. In other words, give Powell more of those possessions, and watch his efficiency drop.Again, this is no knock on Powell or any of the other low-usage, high-efficiency players who populate the league today. The rise of those types — and the dip in correlation between usage and efficiency — is a sign that coaches are actually getting better than ever at identifying players for specific roles that maximize their talents, and not asking them to do more.
A related trend is the increased specialization of players into either high- or low-usage categories. The share of qualified NBA players who either have a usage rate of 15% or lower, or 25% or higher, has risen from 31.4% in 2003-04, two decades ago, to 42.7% in 2023-24:
This is part of what The Athletic’s Seth Partnow (and others) have coined as “Heliocentrism” in the NBA — the trend for teams to have more of their offense revolve around just one or two stars than in previous eras of the game. Powell has the highest offensive rating of any player ever, regardless of usage, in part because Luka Dončić has the second-highest offensive rating of any player with a 35%+ usage rate ever. (Only Joel Embiid, with a ridiculous 126 offensive rating and 38.9% usage this season, is better — and he is arguably the greatest scorer in NBA history.) And when he plays, Kyrie Irving is no slouch as a Mavs teammate, either, with a 121 offensive rating and a 29.1% usage rate.
The league has gotten more efficient because this arrangement — one or two high-usage stars surrounded by a ton of low-usage, hyper-efficient role players — leads to better offensive outcomes than strictly allocating shots so that usage and efficiency match up. It may sound strange that you can improve scoring by giving fewer shots to your most efficient options, but the league has been moving in that direction for decades now — and you can’t argue with the results.
Filed under: NBA