Release and posture in archers: Kuch study

When the string goes, the bow recoils towards the archer and the body has to absorb that push. The study asks whether trained archers expect it, anticipating it with a movement of their own weight, and whether anticipating it sooner helps them score. Thirteen recurve archers, six elite and seven sub-elite, each shot 18 arrows at 70 metres standing with one foot on each of two force plates.

In every shot by every archer, the centre of pressure moves away from the target, with an amplitude between 15 and 20 millimetres: eight to ten times the sway measured during aiming, which was 1.9 millimetres, and four to five times the centre of mass displacement the authors estimate the recoil to cause, about 4 millimetres. In 67 per cent of shots this movement begins before the string leaves, that is before the perturbation that is supposed to cause it.

The elite archers do it more often and sooner. They used the anticipated strategy on 77 per cent of shots against 59 per cent for the sub-elite, and on anticipated shots the movement began on average 23 milliseconds after the clicker fell, about 130 milliseconds before the release, against 119 milliseconds for the sub-elite, which works out at about 30 milliseconds before the release. Twenty-three milliseconds is less than any postural reflex latency, which runs to about 40: the movement was therefore already programmed. On non-anticipated shots the movement began at 163 milliseconds in the elite group and 180 in the sub-elite.

The elite archers also reacted faster to the clicker: 146 against 156 milliseconds. During aiming they swayed less, but on one axis only: the mean and root mean square distance from the mean centre of pressure were smaller in the anterior-posterior direction, that is perpendicular to the shooting line, and in the plane, while along the shooting direction the two groups did not differ.

The part that does not add up is performance. The elite archers scored more, 9.2 against 8.9 per arrow, but shots with an anticipated adjustment were worth no more than shots with a late one (9.0 against 8.9), neither in the overall comparison nor within each group. And once the traces were aligned on the moment the movement starts, the shape of the displacement was identical in both groups and both strategies: what changes is when the body moves, not how.

The elite group was six archers (three men and three women, mean age 27.0, 16.0 years of practice), recruited at the French national elite training centre on two criteria: more than ten years of archery and at least one appearance, in the preceding year, at a senior international outdoor competition (European Championships, World Cup stages, World Championships, Olympic qualifiers or the Olympic Games). The sub-elite group was seven (four men and three women, mean age 16.5, 8.0 years of practice), from a next-generation training centre, with less than ten years of practice and no senior international competition. Anyone injured in the previous six months was excluded.

Each archer did their own competition warm-up for fifteen minutes, including a dozen arrows, then shot 18 arrows on a ten-ring 122 centimetre competition face at 70 metres, with no wind, using their own bow complete with stabilisers and sight and their own arrows. Before each shot the archer stepped onto the two plates, one foot on each, shot at their own pace, then stepped off and started again. The score of each arrow was collected.

The two Kistler plates, sampled a thousand times a second, stood side by side and were oriented so that the medio-lateral axis coincided with the shooting direction and the antero-posterior axis was perpendicular to it. The signals were filtered with a third-order Butterworth low-pass filter at 6 hertz and combined into the global centre of pressure. Every shot was filmed by a high-speed camera at 1,000 frames per second, two metres from the archer and aimed at the riser on the clicker's side; plates and camera were synchronised by an external trigger.

Two instants were identified from the images: the clicker falling, the first frame in which the clicker touches the riser, and the release, the first frame in which the arrow tip moves towards the target. The difference between them is the archer's reaction time. The aiming phase was characterised over the second preceding the clicker's fall, chosen as the shortest duration observed in all participants between full draw and release, using six stability indicators: mean and root mean square distance from the mean centre of pressure, mean and maximum velocity, sway amplitude and the area of the 95 per cent confidence ellipse. The shooting phase runs from the clicker's fall to half a second after the release.

The onset of centre of pressure movement along the shooting axis was defined as the instant, after the clicker's fall, at which sway velocity on that axis crosses a threshold set at two standard deviations above the mean velocity of the aiming phase. If that instant precedes the release, the shot is classed as early, otherwise as late. A total of 230 shots were analysed, 107 from the elite group and 123 from the sub-elite. Since the distributions were not normal, results are reported with medians and interquartile ranges and compared with non-parametric tests: Mann-Whitney U for the aiming indicators and the scores, a two-factor Kruskal-Wallis for reaction time and onset, and Fisher's exact test for the proportion of use of the two strategies. To compare the shape of the trajectories, one-dimensional statistical parametric mapping was used over the half second following the onset.

The limit the authors put first is the number: six elite and seven sub-elite archers, with all the difficulties of scheduling and availability that come with athletes at that level. Only trained archers took part, so the picture does not extend to those who have been shooting a short time, and understanding how this anticipation is learnt would require measuring the same young archers once fully matured, or measuring beginners. Electromyography was not used, a deliberate choice so as not to disturb the shooting sequence, and without it there is no knowing which muscles pre-activate.

The comparison between the two groups does not separate experience from growth. The elite archers averaged 27 years of age and sixteen of practice, the sub-elite 16.5 years of age and eight of practice: they are adolescents, and the authors themselves offer as an explanation, alongside shorter experience, a biological maturation and a body schema not yet complete. Whether training or age makes the difference cannot be said, and neither can which of the two causes the score.

Part of the analysis as constructed remains open to question, and they acknowledge it: the threshold chosen to divide early shots from late ones is the moment of release, a choice that works for the elite group, whose onsets fall into two distinct clusters, but less well for the sub-elite, where early shots begin only 30 milliseconds before the release and late ones 30 after, too close for the distinction to mean much. Read that way, the data suggest that the sub-elite anticipate the release while the elite anticipate the clicker's fall, which is an interpretation after the fact.

Some key figures are not measured here but taken from other work: the recoil kinetic energy of about 0.2 joules, the recoil velocity of 0.26 metres per second and the string and arrow contact time of less than 16 milliseconds. The centre of mass displacement of about 4 millimetres, which serves as the yardstick for the observed amplitude, is an estimate built on those values by treating archer and bow as a rigid body.

Finally the central question goes unanswered, and the authors say so: anticipating is not associated with a better score. With 18 arrows per archer, they add, it cannot be ruled out that a difference exists and is too small to show, given that a qualification round runs to 72, and they cite a model under which 0.2 points per arrow are worth ten places in the ranking. That calculation is taken from another source and applied to a difference that is not significant here.

The most concrete figure here for an archer is the size of the movement the body makes at the release. The centre of pressure shifts towards the heel, away from the target, by 15 to 20 millimetres: eight to ten times the sway of the aiming phase, which was under two millimetres. It happens on every shot by every archer measured, including the best. It is therefore not a failure to hold still: it is part of the movement, and it answers a push the bow gives back when the arrow leaves.

The second thing is that this movement starts before the push, not after. On 67 per cent of shots it begins before the string leaves the fingers, and in the elite group on average only 23 milliseconds after the clicker falls, too short an interval to be a reaction: the body was already going. It is one of the clearest examples of what the automatic side of the shot amounts to, and it shows how finely the clicker is tied to the rest of the movement.

The third, and the one most easily misread, is that anticipating sooner did not score more. The elite archers anticipated more often and earlier than the sub-elite, and they scored more, but within the same archer early and late shots were worth the same. The study therefore offers nothing to apply: there is nothing to anticipate deliberately, and the authors themselves conclude that whether anticipating sooner helps cannot be said.

One detail about aiming is worth noting for its direction. The elite archers swayed less than the sub-elite on the axis perpendicular to the shooting line, the one running forwards and backwards relative to the body, and not on the one pointing at the target. An archer's base of support standing upright is narrower in that direction, and control there falls to the ankle muscles: it is the axis on which, in this sample, the difference in level showed.