The release in archery: anatomy of a few milliseconds

The release is over too quickly to be steered. The most precise measurement comes from sixteen world-class recurve archers filmed at 4,000 frames per second: the string stays on the fingers for 1.9 thousandths of a second on average and the nock leaves the string after 14.6. An older measurement, taken with an accelerometer on the bow, gives twice that, and the gap may come down to where the clock is started; either way we are talking about a hundredth of a second or a little more, roughly a tenth of the time between the clicker going off and the string leaving.

What the measurements have changed most is how we look at the fingers. For decades the release was taught as a relaxation, with fingers that do not open but simply let go, and today the registry rates that rule as refuted: in many strong archers the finger extensor is working, archers who open and archers who relax score the same, and the only three-dimensional measurement of the hand shows an opening that takes place in the joints between the phalanges. Here I follow the release in the order it happens, from the fingers to the arrow and then to the arms, and end with what can be trained, which lies almost entirely before and after that hundredth of a second.

What happens, millisecond by millisecond

The best-documented sequence comes from that footage, filmed at the London Olympic test event in 2011 on eight men and eight women shooting at 70 metres. The times are means, counted from the instant the archer starts to release, and they change little from one archer to the next: standard deviations are no larger than 1.3 thousandths, and between men and women, with different draw weights and arrows, the difference stays below one and a half thousandths.

1. Phase 01 · 0-1.9 ms

   The string leaves the fingers

   The string slips away in 1.9 thousandths on average, between 1.50 and 2.25 across the sixteen archers. In 69% of cases the last finger to let go is the middle finger, and in 75% the first is the ring finger. In that time the nock and string move sideways by 4.7 millimetres on average, and it is this sliding across the fingertips that starts the arrow flexing.

2. Phase 02 · 1.9-8.6 ms

   The arrow bends and straightens

   Once clear of the fingers, the string keeps moving towards the body and the nock reaches its closest point at 5.4 thousandths. Contact with the pressure button ends between 5.5 and 8 thousandths, and at 8.6 the arrow is straight again. This is the flexing of the archer's paradox, and it is what lets the rear of the arrow get past the grip without touching it.

3. Phase 03 · 8.6-14.6 ms

   The nock leaves the string

   The arrow flexes the other way, peaking at 12.4 thousandths, and leaves the string at 14.6 while still bent. Up to that moment, in none of the sixteen archers did the bow move sideways by more than a millimetre or rotate by more than a tenth of a degree.

4. Phase 04 · 14.6-18 ms

   The arrow passes the bow

   In the archer shown frame by frame, the shaft passes closest to the button at 15.5 thousandths, about 150 millimetres from the nock, so after it has left the string. The nock clears the button at 18 thousandths, with room to spare for the vanes as well.

Scroll the figure→

On the overall figure, though, the registry rates the number as contested. In 1987 an accelerometer mounted on the bow of an Italian national team archer gave about thirty thousandths, three hundredths of a second, from the start of the limbs' return to the arrow leaving the string, and in 1994 another Italian group repeated the same figure without measuring it. It is double the London footage, but its zero point was taken from the clicker and the string used as electrical switches, which may not coincide with the start of the release, and the two measurements have never been compared on the same shots. The order of magnitude, a hundredth of a second, is the same in every study.

This has a practical consequence for anyone filming themselves on a phone. At 240 frames per second each frame covers a little over four thousandths: the string leaves the fingers within a single frame, and from the start of the release to the arrow leaving the string there are three or four. Slow-motion video is therefore for seeing what the hand and arm do before and after, not how the fingers let go of the string.

The fingers: they open, and where

The rule of the passive release says the fingers do not open: they relax, and the string goes by itself. Seen from outside, it holds up. In competition video of thirty elite archers, twenty-eight let go of the string with their fingers still curled, and in the London footage the fingers do not straighten. The visible shape of the hand, however, says nothing about which muscles are working, and under the skin the picture is different.

The study that carries most weight dates from 1990: fifteen highly skilled archers, nearly all from the US national team, with electrodes on the finger flexor and extensor. Eight let the activity of both muscles fall away, seven had a brief burst of the extensors in the 200 thousandths before the release, and each repeated their own pattern every time; the season's mean score was practically the same, 1,209 for those who relaxed and 1,213 for those who opened, out of 1,440. In a laboratory in Ankara, with ten elite Turkish national team archers, ten beginners and ten people who had never shot, the extensor contracted after the clicker in all three groups while the flexor gave way gradually, and only one archer, among the top ten in the world, showed the drop in the extensor that the rule would prescribe. So the rule does not describe how everyone releases, and above all it does not separate those who shoot better: fingers that stay curled are an effect of the release, not the instruction that produces it.

Where the fingers open has been measured by a single study, with eight infrared cameras at 500 frames per second on an archer from the Austrian B squad, 56 arrows at 18 metres. The knuckles barely move, between 6 and 9 degrees, and almost all of the opening happens in the middle and end joints: 81 degrees in total for the index finger, 58 for the middle, 48 for the ring finger. That is also why the hand looks closed from outside, since the base of the fingers hardly moves. In the same shots, how fast the middle finger opened did not separate arrows in the X from those scoring 8 or less, so opening faster did not bring the arrow closer to the centre. Going back over the same arrows, the authors did find a difference in regularity: on the poor arrows the curve of the joint at the base of the middle finger strayed further from the group average. It is one archer, and it serves as a pointer to what to look at, consistency rather than speed.

Before the release the question changes: on better arrows, is the string hand already more relaxed while aiming? There are two answers, and they point in opposite directions. In four elite Turkish women archers, shooting indoors at 18 metres, the finger flexor in the second before the clicker was at 80.0% of maximal contraction on arrows scoring 10 and at 85.6% on those scoring 8; in eight high-level French archers, over 549 arrows at 70 metres, both flexor and extensor were instead more active on the 10s. Distance, level and time window differ, and neither study is enough to dismiss the other: a "soft" hand while aiming remains a technical preference, not a measured result.

One detail comes from a Japanese group in 1987. In the most skilled group of twelve archers, at 8 metres, the deltoid fell silent for an instant in 70-100% of shots, about sixty thousandths before the release, and the finger extensor began to work at exactly that moment; in one archer filmed at 200 frames per second the fingers started to open 25 thousandths before the string moved off. In the other groups the silence appeared less often. It is a single research group, with no statistics across levels and no link with the score of individual arrows, but it points the same way as the studies with electrodes on the fingers: in the more skilled archers the extensor comes into play right at the release.

What the arrow does, and why the fingers matter

The arrow's flex starts with the way the string leaves the fingers. As the grip gives, the string slides across the fingertips sideways, towards the body, and forwards: 4.7 millimetres sideways on average in the London footage, and in the order of 9 or 10 forwards. With a mechanical release that sliding is missing, and the difference shows: in a Japanese experiment with the same competition arrows, the mean amplitude of flex just after leaving the bow was 1.22 millimetres with a release aid mounted on a robot with a recurve bow and 6.86 when two national-level archers shot them, more than five times as much. The vibration frequency stayed at around 86 hertz, so what changes is how far the arrow bends, not the way it vibrates.

It would be tempting to conclude that less flex means better shots, and here the measurements say something less obvious. In the same experiment drag was more consistent from shot to shot precisely on the arrows shot by the archers: the standard deviation of the drag coefficient was 0.073, against 0.103 with a mechanism imitating the fingers, 0.119 with the release aid and 0.267 with a compressed-air launcher that does not make the arrow flex. The authors infer that flex of the right amplitude makes the boundary layer around the shaft turbulent, and with it the flight more repeatable. It is a single experiment, with few launches per condition, and the link with a tighter group on the target has not been measured; it is enough, though, to say that the flex the fingers give the arrow is not simply a fault to be reduced.

Finger speed does not behave as expected either. A shooting robot with motorised fingers, built in Japan in 2015, opened its grip at three different speeds, and the arrow's maximum flex came out at 7.07, 7.44 and 7.23 millimetres: it did not grow with speed, and the peak was at the middle one. The differences are a few tenths of a millimetre, on a machine and without statistical tests, but they point the same way as the Austrian archer: neither result points to finger opening speed as something to work on.

That leaves the passage past the bow. The common assumption is that the critical point is the rear of the arrow, with the vanes brushing the button; in the London footage, however, the point where the shaft passes closest to the bow is about 150 millimetres from the nock, and the passage happens once the arrow has already left the string. When the rear reaches the button the arrow has only just begun to spin, so the vanes go past oriented as they were at full draw. The clearance with which the rear passes the button, on the other hand, varies widely from archer to archer, from about 2 to 25 millimetres, and the authors link it to the way the bow hand torques the grip: in tuning, vane clearance has to be checked on your own shot, not inferred from someone else's.

The arms and the hand after the release

While the arrow is on the string, in high-level archers the bow stays still: in the sixteen London archers it moved sideways by no more than a millimetre, which is the resolution of the camera. A Korean study from 2007, at a thousand frames per second, found the same in the then world number two, with zero displacement, while a university archer's bow moved by one or two tenths of a millimetre. The bow's large movement comes after the arrow has gone.

After the release the arms move apart. The only study to have measured it, with an accelerometer on each arm of two university recurve archers at 70 metres, found that the two arms move apart in opposite directions. Vertically, the bow arm drops and the string arm rises by similar amounts, around 2-3 centimetres; along the line of the shot, though, the bow goes forward much further than the hand goes back, 6.8 against 1.2 centimetres on low-scoring arrows and 9.2 against 1.6 on high-scoring ones. Two archers and no statistical test: the shape of the movement is visible, a link with the score is not.

The string hand slides back rather than staying put. In 1976 a Japanese group photographed it in four experienced archers, while in six complete beginners the hand stayed where it was; in 2007 the thousand-frame camera measured, in the few thousandths between the release and the arrow leaving, a forward-backward displacement of 1.24 millimetres in the world number two and 0.70 in the university archer. In both comparisons the movement is more marked in the better archer. How far the hand should travel back over the following tenths of a second, and whether that changes the score, has not been measured across several archers of the same level.

What the hand does next is the follow-through, and here too it pays to separate what can be seen from what is known. In better archers the shoulder and back muscles stay active after the release, and in two world-class archers the activity lasted about a second and a half; whether arrows shot with those muscles active land closer to the centre, however, is contested, because alongside one supporting study on ten club archers there are two studies at the level of the single arrow that point the other way. On the range, a hand that comes forward towards the bow is read as a sign that the draw stopped before the release, and a hand that flies out sideways, away from the neck, as a release pulled away with the arm: these are coaches' readings, not measurements, but they are the ones the follow-through lets you make by eye.

What separates experts from the rest

Put side by side, the measurements comparing different levels all speak of repeatability and of where the work is done, not of a faster or "softer" release.

What you can train, if the release cannot be steered

If nothing can be corrected within that hundredth of a second, the work goes into how you arrive at the release and how you come out of it. There are four conditions you can actually influence.

1. An identical anchor

   The anchor point is where the string starts from: if it shifts from one arrow to the next, so does the start of everything that happens in the following thousandths.

2. Expansion through the back

   In better archers, while aiming, the work moves from the arms to the shoulder blades. In thirty Japanese archers the elite used, on the string arm, less biceps than the other two groups, less triceps than the beginners and more lower trapezius, 54% of maximal contraction against 39 and 34 in the two weaker groups. The idea that elite archers activate every muscle less is contested; what holds up better is the shift of work towards the shoulder blade.

3. A steady gaze

   In one study of British archers, the Olympic-level group held their gaze still on the target for almost two seconds before the release, against less than one in beginners, and the fixation started earlier. That a longer quiet eye improves accuracy rests for now on that single comparison between levels.

4. The clicker, and the time that follows it

   Between the clicker going off and the string leaving, between one and a half and two tenths of a second pass, and among elite archers it seems to matter more that this time stays the same. In seven German national team archers, those who kept it most consistent had the highest averages, while in the largest sample, thirty-nine elite archers, a quicker reaction accounted for barely half a point. Whether the release is a reaction to the click or a movement already under way when the click arrives, the only direct comparison, on nine German national team archers, tips the balance towards the second reading.

On that last point there is one more piece of data. Nine elite women recurve archers, at 8 metres, shot half their arrows with the clicker and half without, releasing when they decided for themselves, and without the clicker their aim at the moment of release was more tightly grouped; in the laboratory, eight people with no shooting experience using a simulated bow grouped their shots better when they chose the moment than when they waited for a visual signal, and waiting for the signal stiffened the ankle. This is not a case for removing the clicker, because the women's accuracy was calculated from the bow's orientation rather than from arrow impacts, and arrow scores were not measured. It does suggest that reacting to an external signal is not necessarily the steadiest way to let the shot go.

Three exercises for a repeatable release

The exercises below come from coaching practice, not from studies that have measured their effect: they are there to isolate the conditions above, one at a time.

• Exercise 01

  Blank boss with eyes closed

  A few metres from a boss with no target face, shoot short sets with your eyes closed after checking the anchor. Taking away the aiming leaves your attention on the feel of the draw and the release, the grip that stays the same and the hand that slides back along the neck. It is for building the movement, so the quality of each arrow matters more than the number.

  Removes

  Aiming

  Trains

  The feel of the release

• Exercise 02

  Holding at full draw without shooting

  Come to anchor, keep expanding for a few seconds beyond your normal aiming time, then let down without shooting. It works on holding the shoulder blades, and it separates aiming from the moment of release, which helps anyone who feels they have to shoot within a set time.

  Trains

  Expansion

  Note

  A draw weight you hold comfortably

• Exercise 03

  Clicker with the draw still moving

  With a recurve, the exercise is to bring the clicker off through back expansion, not by pulling with the arm, and to let the release come with no pause between click and string. The time between the two can be read from the sound, recorded on a phone, and what to watch is how much it changes from one arrow to the next rather than how short it is. With a compound the same principle applies to the release aid: the shot goes while the draw keeps pulling into the wall, not through a separate movement of the finger.

  Trains

  Continuity

  Measures

  Consistency of timing

Recurve and compound

In recurve and barebow the release belongs to the fingers, with everything described above: the string sliding sideways across the fingertips, the arrow flexing, the clearance past the bow that also depends on each archer's shot. In compound the string is held by a release aid, whether trigger, thumb button or back tension, and the arrow flexes far less: with a release aid on a recurve, the Japanese experiment found about a fifth of the flex the archers produced, and for a compound a model gives around 2 millimetres. On the archer's side the issue is the same. With a release aid the fault to guard against is firing the shot with a movement of its own, punching the trigger or flinching ahead of it: it is the equivalent of a release snatched with the fingers on a recurve.

When the release gets away from you

The release is automatic in the sense that nobody steers it thousandth by thousandth, and as the shot is learnt it changes in the brain as well. Magnetic resonance studies find a busier brain in beginners, especially in the frontal areas, and a seven-week course tracked with electroencephalography and electromyography shows less information flowing from the cortex to the muscles by the end; among archers who are already trained, though, the only comparison goes the other way, with elite women archers activating some motor areas more than university archers do. That is why the idea that the shot comes under less cortical control is contested: it holds for the first step of learning, and beyond that step it has no support.

When control breaks down, the release can go before its time, and archers with an early release describe it like this: the arrow goes off by itself, even when they try to hold it. The account comes from very few people, almost all of them kyudo practitioners, and "against their will" only partly holds: of the three practitioners with an early release examined in a Japanese case series, one could partly hold it back, and in none of them did the problem always show up the same way. Outside kyudo the closest account is that of an English club archer, who could not say whether the fingers let go of the string by choice or because they gave way.

This is target panic territory. On its origin the registry holds definitions and accounts, not measurements. I have put forward a hypothesis, that the problem arises when the release loses its automaticity, and set it out in the Lab with seven testable predictions: it is a hypothesis awaiting its outcome, not a finding.

One caution remains on attention. The idea that attention should go outside the body rather than to the movement comes mainly from other sports, and in archers the only trial to compare the two, gaze on the sight against the feel of the body, found no difference in score, with six arrows per condition, too few to pick up a small effect. In a movement that lasts a hundredth of a second, at any rate, thinking about your fingers while you shoot has little room: work on the fingers belongs in the exercises, and in competition attention goes wherever the focus each archer has built points it.

The release as a consequence

Taken together, the picture is fairly clear. The release lasts a hundredth of a second or a little more, too little to be steered, and the fingers open at the joints between the phalanges, with the same scores for those who relax and those who open. The arrow flexes because of the way the string leaves the fingers, and that flex has a useful side too. What separates the levels, in the measurements, is how repeatable the movement is and how the work moves from the arms to the shoulder blades.

For anyone who wants to improve their release, the order I recommend follows the conditions: first the anchor and expansion, until reaching the clicker is the same movement every time, then the release exercises, and only at the end competition work. Copying a champion's release by eye does little good, because the outward form, the fingers that stay curled and the hand that slides back, is the result of everything that comes before.

@misc{campagna2026_rilasciotiroconarco,
  author = {Campagna, Matteo},
  title = {The release in archery: anatomy of a few milliseconds},
  year = {2026},
  howpublished = {matteocampagna.com},
  url = {https://matteocampagna.com/en/articoli/release-in-archery},
  note = {accessed 2026-10-16}
}