Usage Rate Is Zero-Sum: Five Players Always Add to a Hundred

Short stints inflate it beyond meaning

The same eighteen shots taken in eighteen minutes rather than thirty-six read as 58.5 per cent usage, which is not a level any player sustains over a season and is not what the number is claiming. It is a rate, so a small denominator makes it explode, and a substitute who takes four quick shots in a three-minute burst can post a figure that would top any leaderboard. Sample size is the whole difference between a usage rate that describes a role and one that describes a fluke: below a few hundred minutes the figure belongs in a footnote, and comparing a starter's season to a bench player's handful of appearances is comparing two different kinds of measurement.

Pair it with efficiency or it says nothing

Two players at 29 per cent usage can be having opposite seasons, and the pair of numbers that separates them is usage against true shooting. High usage with high efficiency is the rarest and most valuable combination in the sport, because efficiency usually falls as volume rises — defences adjust, shots get harder, and the easy attempts are used up first. High usage with poor efficiency is a player taking possessions his teammates would have used better, which is a coaching problem rather than a statistic. Low usage with high efficiency is often just someone taking only open shots, which is valuable and does not scale. Read the two together or neither means much.

Frequently asked questions

Why are free throws worth 0.44?

Because a free throw attempt is not a possession. Most fouls send a player to the line for two shots, so a single attempt is roughly half a possession; the coefficient is trimmed below 0.5 to account for and-ones, three-shot fouls and technical free throws, which end differently or do not end a possession at all. The 0.44 is an empirical average fitted to real play-by-play data, not a derivation, and its exact value drifts a little with the rules and the era. Nothing in a usage comparison turns on the third decimal, but a formula that counted every free throw as a full possession would badly overstate anyone who lives at the line.

Should a team spread usage evenly?

Not necessarily, and the flat 20 per cent that comes out of a perfectly even split is a mathematical curiosity rather than a target. Concentrating possessions in the hands of the player who does most with them is usually right, up to the point where the defence can plan around one person and the efficiency starts to fall. Where that point sits is a question about a specific squad, not about the statistic, which is exactly why usage is a description rather than an instruction. The one thing the zero-sum property does tell you is that raising anyone's usage lowers somebody else's by the same amount.