True shooting is a poor way to judge a role player

One number that flattens volume, difficulty and gravity into a single percentage will always lie about the men at the edges of a rotation.

True shooting percentage did enormous good. It killed the era of judging scorers by field goal percentage, which punished shooters and rewarded anyone tall enough to dunk. Weighting threes properly and folding in free throws made the number honest about what a possession is worth. As a first pass on a high-volume scorer it's still the best single stat we have.

As a way to evaluate the eighth man, it's close to useless.

The problem is that it treats every attempt as an independent event with no context, and role players live entirely in context. A wing who shoots thirty-eight per cent from the corner but does it on catches that arrive with a defender two feet away is doing something completely different from a wing who shoots forty per cent on wide-open kickouts generated by somebody else's gravity. Same output. Different jobs, different difficulty, and the difficult one is more valuable because it survives playoff defences.

Then there's the shooter who doesn't shoot. A guy standing in the corner at forty per cent on two attempts a game is applying constant pressure to a defence he never touches. His true shooting is built on a sample so small it's noise, but the low man honouring him is why the roller gets a clean lane. The value shows up in somebody else's line. Meanwhile the volume-hunting bench scorer who takes eleven contested shots and lands at fifty-three per cent true shooting will look comparable in a spreadsheet while actively damaging possessions.

And the metric punishes the specific things you want a role player to do. Attack a closeout hard, draw the second defender, get fouled without a shot going up. That's a good possession with no shot attempt in it. Dive on a roll and finish through a body at fifty-eight per cent from three feet, which will drag the number down relative to standing outside and firing. Take the tough end-of-clock shot because the star has been doubled. All of that is punished.

If you're actually evaluating a bench player, watch how many of his makes come from shots the defence chose to give him, and how many defensive attentions he draws while standing still. Those two things tell you almost everything, and neither of them appears in the number that gets quoted in every broadcast graphic.

Efficiency metrics are answers to the question of who scored well. They are not answers to the question of who helped.