One trial is not a measurement of stability

Stability is variability, so recording a single repetition throws away the only part of the signal worth reading.

Ask what a stability score means and you usually get a number taken from one repetition of one task on one day. A single hop. A single landing. Somebody wrote it down, somebody else compared it to a threshold, and a decision followed.

But stability is not a value. It is a spread. A knee that lands in the same place five times running is a stable knee; a knee that lands in five different places, one of which happens to be textbook, is not stable even though its best trial looks excellent. Record one repetition and you have recorded a point drawn from a distribution whose width you never measured, and the width is the whole subject.

This matters in the direction that flatters people. Give an athlete three attempts and keep the best, which is standard practice almost everywhere, and you have systematically selected the tail of their distribution. Their typical landing is worse than the number in your file, and it is worse by an amount that grows with how inconsistent they are. The least stable athletes get the most generous flattering, because they have the widest spread to select from. That bias points exactly the wrong way.

I would rather have five untidy trials than one clean one. Take the mean and take the range, and note that the range is often the more interesting column. Two athletes can score identically on the average and be describing completely different knees.

There is also the fatigue question, which people fold into the same session and should not. Trial one and trial five differ partly because of noise and partly because of tiring, and a five-trial protocol cannot separate the two. If the last landing is always the worst, that is a drift, not a spread, and it wants a different label and a different response.

None of this needs new equipment. It needs the recording sheet to have more than one box on it, and it needs whoever built the threshold to say which trial the threshold came from. Most published cut-offs are silent on that, which means a squad recording means is comparing itself against a table built on bests.

Those two numbers can differ by more than the effect anyone is trying to detect.