The contested shot label counts hands that arrived too late to matter

Tracking systems record whether a defender was near the shooter, which is not the same as whether the defender affected the shot.

Two defenders end up two feet from a shooter at the moment of release. One got there while the ball was still coming up and the shooter had to adjust the release point. The other arrived flying past, hand up, a beat after the ball had left. The box score of the tracking system logs both as contests.

That is the whole problem in one sentence, and it has quietly polluted a lot of defensive evaluation.

The systems generally define a contest by proximity at or near release, sometimes with a rule about the position of the defender's hand. It is a sensible thing to compute because it is computable. What actually degrades a shot, though, is not presence. It is whether the shooter had to change something. Whether the release got rushed, whether the feet were set, whether the eyes came off the rim, whether the arc had to go higher than the shooter wanted.

Timing decides that, and timing is exactly what proximity at release cannot capture.

This produces a specific kind of bad defender who grades well. He is fast, he is late, and he covers ground spectacularly in the final half second. Every camera loves him. The tracking data loves him too, because it counts his arrivals. What the shooter experiences is a defender who was not there when the decision was made and turned up afterwards, which is close to no defence at all and occasionally worse than none, because a flying closeout that arrives late is an invitation to shot fake and drive.

Meanwhile a defender who is simply in position early, hand up, feet quiet, never gets the ball swung to his man at all. He generates nothing. No contests, no deflections, no stat.

If I were building the metric I would care about the defender's distance and closing speed at the moment the shooter caught the ball, not at release. That is the moment the decision to shoot gets made. Everything after it is theatre.

Some teams do this internally. Very little of it reaches public analysis, which is why so much of the online argument about wing defenders runs on numbers that reward the wrong behaviour.

Watch a possession with the sound off and ask a single question: at the moment he caught it, did the shooter feel like he had time? That answer is the whole of the thing, and no current public metric gets near it.