Free throw percentage is a weak alibi for a broken jumper

Draft models lean on the foul line because the sample is clean, and that convenience has cost teams real money.

Every draft season somebody defends a prospect who cannot shoot by pointing at the foul line. He shoots eighty percent from there, the argument goes, so the jumper will come. I have heard it about a dozen players and I can think of very few for whom it turned out true in the way the argument promised.

The appeal is obvious. Free throws are the cleanest shooting sample a model can get. Same distance every time, no defender, no clock, no shot preparation, and enough attempts in a college season to mean something statistically. Compare that with three point attempts, where a player might take ninety in a year off different footwork against different closeouts, and you can see why a model reaches for the tidy number.

But the foul line measures the wrong half of the skill. Shooting off the catch is a sequence. Read the pass, set the feet, load, release, all under time pressure and often while falling away from a recovering defender. A free throw removes every part of that except the release. So the statistic tells you the hand works. It says almost nothing about whether the player can get into the shot at speed, which is the part that actually fails at the professional level.

There is a subtler problem. A high foul line number on low volume tells you a player rarely got to the line, which usually means he rarely attacked, which is itself information the model throws away by treating percentage and attempts as separate columns.

What I would weight instead is unglamorous. Attempts off movement. Whether a player shoots the same way in the fourth quarter as the first. Whether his misses are short or long, because a consistently short miss is a strength problem, which improves, while a scattered miss is a mechanics problem, which often does not. None of that is in a public box score, which is exactly why the teams with real video departments should have an edge and frequently do not use it.

I am not saying ignore the number. A prospect who shoots sixty percent from the line has told you something bad and you should believe him. I am saying the good version of that number is not the promise scouts treat it as, and the gap between eighty percent standing still and thirty five percent moving is where a lot of first round picks quietly disappear.