One thing first, because it matters: what follows is my hypothesis, not settled science. It draws on the neuroscience of movement, it makes predictions you can test, and it has weak spots, which I lay out. I put it forward because it is the clearest way I have found to fit together pieces that, on their own, refuse to add up, and alongside each step I say what the studies show.
Calling it mental explains only part of it
The head certainly plays a part, but stopping there leaves three questions unanswered: why it is the release that goes, and not the draw or the aim; why the problem so often arrives all at once, from one day to the next, rather than creeping in slowly; and why the symptoms vanish in front of a blank bale, only to return when the stakes go up.
A purely mental account predicts none of this, and neither does a purely physical one: if a muscle or a nerve were giving way, the trouble would not switch on and off with the size of the stakes. Yet in archers' own accounts that is exactly what it does: it vanishes when the target does not matter and returns when it does. The problem sits squarely in between, part body and part head, and that is where to look for it.
Nor, in archers, does one of the most common versions of the mental explanation hold up well: that target panic strikes those who fear other people's judgement. The only survey to test it brings together 69 archers and 86 golfers, all British competitors: across the whole sample those who have had the problem report more social anxiety, but among the archers social anxiety is the same with target panic as without, 3.07 against 3.05, and the difference comes from the golfers with the yips. The need not to display one's flaws does point the same way in the archers too, but the gap is small. That is why, in the registry, the claim that those who have had the yips report more social anxiety and a stronger need to hide their flaws is contested.
Two ways to run a movement
It starts from a simple fact: a movement can be run in two very different ways, and the difference between them is decisive. In the first, the command is worked out in advance and fired off without waiting for the senses to report back; engineers call it open-loop control, I call it automatic. In the second, the movement is watched as it unfolds, checked against what the eyes and body report, and corrected on the fly; that is feedback, or closed-loop, control.
That practice makes a movement need less and less watching is the idea everything rests on, and in part it shows up in archers' brains. Three Korean MRI studies of archers simulating or imagining the aim find a more active brain in beginners, mainly in the frontal areas, and a narrower network in experts. In twenty-six first-year students followed through a seven-week course, less information flows from the cortex to the shooting muscles by the end, which the registry records as reported, because there is one study and no control group. The picture holds only for the first step, though, from not knowing how to shoot to knowing how. Among trained archers the only comparison points the other way, with elite women archers who had won Olympic and world medals activating some motor and frontal areas of the cortex more than university archers who had been shooting for almost twelve years; and the authors of the study on the students read the drop as more inhibitory control from above, not as the cortex stepping aside. That is why the claim that as the shot is learnt, it comes under less cortical control is contested in the registry, and nobody has yet measured what happens in the brain of an archer who loses the movement.
Aiming and holding lean on feedback the whole time, and rightly so, because they last long enough to be corrected. The release, by contrast, is over too quickly.
The release is too fast to be watched
It comes down to the numbers. In high-speed footage of sixteen world-class recurve archers, the string leaves the fingers in about two thousandths of a second, and the arrow leaves the string after about fifteen. An older measurement, on a single archer and perhaps from a different starting point, gives twice that, so in the registry the exact figure in the claim that about fifteen thousandths of a second pass between the start of the release and the arrow leaving the string is contested; the order of magnitude, a hundredth of a second, is the same for everyone.
Reacting to a signal takes time of a different order. Between the clicker going and the arrow leaving, a trained archer takes between one and a half and two tenths of a second, around ten times as long as the release itself. That length is consistent with a simple reaction to a sound, although on its own it cannot say whether that is what fills it or the tail end of a movement already under way. A correction that starts from the eye has to notice the error, decide and send a new command to the muscles, so by the time it was ready the arrow would be long gone: there is no way to watch the release as it happens and fix it. The release therefore cannot be steered while it happens: it has to go off right from the start, which may be why the best releases seem to happen by themselves.
My hypothesis: the release de-automatizes
The idea is that under pressure, with the fear of blowing the shot that matters, the release regresses: it drops out of the automatic, soft control that years of practice had built, and back into the watchful, stiff control the motor system reaches for whenever something feels shaky. But that watchful mode is far too slow for a movement lasting a hundredth of a second, so much so that a ballistic action forced under conscious control can only come out wrecked: frozen, snatched, plucked. Target panic, on this view, is the de-automatization of the release.
The crux of this reading is that the trouble comes from a control mode that is healthy in itself but switched on where it is not needed, rather than from a broken part. It is the same control used to learn a new movement, the right one for a beginner, which on an expert's release becomes an obstacle.
It is also how several archers describe their own problem. Seven British archers with target panic, from club to national level, interviewed at length, all trace its onset to the moment they began thinking about the score of every arrow, and describe what followed as a need to take full control of the shot, which squeezed out the automatic reflexes the shot had relied on. The claim that archers with target panic trace its onset to focusing on the outcome and on expectations is reported in the registry, and it is the description that comes closest to my idea; but it is the memory of people who lived through it, not a measurement, and in the same interviews some point to other origins.
Nobody has yet measured what happens at the release in archers who suffer from target panic. That is why the claim that target panic arises when the release loses its automaticity has the status "predicted" in the registry: in the essay I set out the predictions that would test it, and they are still waiting for an outcome. In 2026 a theoretical model from Ukraine, independent of my work and likewise without data, proposed a similar mechanism: it is another hypothesis, not a confirmation.
Why this reading holds together
This view answers the three earlier questions one by one. Why only the release: because it is the one truly ballistic movement in the shot, the only one that cannot survive being watched. Why all at once: because a stiffness that climbs slowly can cross a threshold and tip a clean release into a freeze from one shot to the next; the collapse is sudden even though, underneath, the cause is building smoothly. That is also how it was described by those who went through it at an English university club followed for a year: a problem that comes without warning rather than as a gradual decline, and that for one archer came straight back as soon as someone pointed it out. Why blank bale brings relief: shooting up close, with no target face and no score, strips away the outcome that switches the monitoring on, and gives the movement room to become automatic again. Whether that relief carries over to the target is another matter: in the few accounts gathered so far, the symptoms came back for several archers as soon as the target face went up again, and in the registry the claim that shooting at a blank bale cures target panic is contradicted.
The same reading makes a prediction about one of the most common pieces of advice, to relax. If the trouble lies in the way the movement is run, lowering tension should not be enough, because it leaves control where it is; what should help instead is whatever moves control away from the movement, such as a neutral target or attention directed outside the body. It is one of the seven predictions in the essay, and like the others it is still waiting to be tested.
Finally, it settles the old body-versus-head argument. On this view the two are not rivals: pressure (the head) makes the control regress, and the stiffened movement (the body) follows directly from it, in a chain of cause and effect.
Where I could be mistaken
I hold this as a hypothesis, not a verdict: in the essay I set out seven predictions that would sink it if they came out the other way. For instance: the antagonist co-contraction during the release should scale with the stakes, gradually, and stay specific to the release; and the physiology underneath should shift smoothly while the visible movement flips sharply at a threshold. I am putting it in writing precisely so that someone can try to take it apart.
There is also a finding the hypothesis has to reckon with, and it comes from kyudo, Japanese archery, which has one name for the release block, motare, and another for early release, hayake. In a small series with electromyography, shooting a rubber bow, the three practitioners with the block had a stereotyped, task-specific problem that eased in two cases out of three when the instructor touched the arm or shoulder, and in one of the two who had it during the recording a co-contraction of the forearm flexors and extensors appeared, starting about a second before kai and thinning out about five seconds into kai until it disappeared, and only then did the release come. These are typical marks of a task-specific focal dystonia, and in the registry the claim that in kyudo the release block resembles a task-specific focal dystonia is reported, with a caveat that matters: there are six cases in two studies, and the resemblance is described, not diagnosed. Early release, examined in the same way, was task-specific too, but it was not stereotyped, did not ease with touch and showed no co-contraction, and one of the three could partly hold it back by will.
For my hypothesis this is a serious question. I read the freeze and the snatch as two outcomes of the same change in control, whereas in kyudo the two forms seem to leave different marks: if larger studies confirmed the difference, the hypothesis would have to explain it, or apply to one form only. The comparison with the Japanese studies, problem by problem, is in the article on premature release and the yips.