True shooting percentage, and what field goal percentage hides
Statistics
Why the most quoted shooting number undercounts three-point shooters and free throw merchants, and what to use instead.
Field goal percentage counts makes divided by attempts and treats every make as equal. A two-foot layup and a twenty-six-foot three both count as one. Since the second is worth half again as much, the number is wrong in a specific, predictable direction: it flatters players who shoot close to the rim and punishes those who shoot from distance.
A player who makes 40 per cent of his threes produces 1.2 points per attempt. A player who makes 55 per cent of his layups produces 1.1. Field goal percentage says the second man is far better. Points per shot says the first one is.
True shooting percentage fixes both problems at once. It counts total points scored, including free throws, and divides by an estimate of total scoring attempts, with free throws weighted as fractions of a possession because two of them usually come from one trip. The result is on the same scale as field goal percentage, so it reads familiarly, but it is actually measuring efficiency rather than accuracy.
The free throw part matters more than people expect. A player who draws contact constantly generates points on possessions that never show up as field goal attempts at all. Under the older number he looks like an inefficient scorer. Under true shooting he looks like what he is.
What true shooting does not capture is equally important. It says nothing about shot difficulty, so a player who only takes wide open shots created by others will outscore, on this measure, a player who creates his own against a set defence. It says nothing about turnovers, which end possessions with zero points. And it is heavily sample dependent early in a season.
If you want two numbers rather than one, pair true shooting with usage rate, which estimates the share of a team's possessions a player finishes. High efficiency on low usage is a specialist. Moderate efficiency on very high usage is often more valuable than it looks, because somebody has to take the hard shots when the clock is running out.