Balance, strength and archery scores: Fan 2025 study

Thirty-seven Taiwanese national-level archers, 19 recurve and 18 compound, each shot six arrows on an outdoor range at their own competition distance (70 metres for recurve, 50 for compound), standing with the left foot on a force plate and wearing inertial sensors on the left ankle, the lower back and both wrists. Separately, a handheld dynamometer measured the strength of four shoulder and scapular muscle groups on both sides.

The recurve archers came out stronger on four of the eight strength measures: shoulder horizontal abduction on the drawing side (19.0 against 15.4 kilograms force), horizontal adduction on the same side (15.5 against 12.6), lower trapezius on the bow side (6.5 against 5.1) and horizontal abduction on the bow side (19.1 against 15.4). The upper trapezius separated the groups on neither side. The compound archers scored more per arrow (9.1 against 8.3), but they shot at a shorter distance on a smaller face, so the two numbers cannot be set against each other.

On stability the picture splits between before and after the release. In the half second before the release the recurve archers had a smaller total centre of pressure displacement (51.9 against 61.6 millimetres) and a smaller sideways sway (1.1 against 1.5 millimetres), while the forward and backward sway did not differ. In the tenth of a second after the release the total displacement was similar again (13.5 against 14.0 millimetres), but the forward and backward sway was much larger in recurve (5.4 against 3.2 millimetres). After the release the angular velocities of the segments were lower in the compound archers almost everywhere: left ankle 2.2 against 3.9 degrees per second, bow-side wrist 104.7 against 132.6, drawing-side wrist 215.5 against 271.4. The lower back was the only part where the difference did not reach significance (19.8 against 29.3). Before the release all these velocities were very small, between 1.7 and 12.4 degrees per second, and only the drawing wrist differed (3.1 against 7.5).

The regressions, adjusted for sex, age, weight and strength, identify different factors for the two bows. In recurve the score per arrow was predicted by total centre of pressure displacement and by the angular velocity of the bow-side wrist, both in the tenth of a second after the release, with standardised coefficients of minus 0.351 and minus 0.259. In compound it was predicted by the forward and backward sway of the centre of pressure and by the angular velocity of the drawing-side wrist, both in the half second before the release, with coefficients of minus 0.242 and minus 0.259. In all four cases smaller values go with higher scores.

The participants came from high school and university teams in Taiwan, recruited between January and July 2024, aged 15 to 24, with at least three years of structured training: the authors call them pre-elite, athletes competing regularly at national level but without international experience. All were right-handed, that is drawing with the right hand. Anyone with a history of shoulder surgery, a shoulder muscle disease or impingement syndrome was excluded. The qualification scores closest to the test date averaged 550 for high school recurve, 624 for university recurve, 663 for high school compound and 670 for university compound.

Testing took place on a regulation competition range under controlled weather: temperature between 26 and 33 degrees, relative humidity between 75 and 80 per cent, no measurable wind. After ten minutes of standardised warm-up, each archer shot six arrows on a five-colour face with ten rings: 122 centimetres with a 12.2 centimetre ten for recurve at 70 metres, 80 centimetres with an 8 centimetre ten for compound at 50 metres. After the postural test came thirty minutes of rest before the strength measurements.

The kinetics come from an AMTI force plate under the left foot at 1,000 samples per second, from which centre of pressure displacement is derived, split into a medial-lateral component (across the target face) and an anterior-posterior one (in line with the face). The angular velocities come from Blue Trident inertial sensors at 1,000 samples per second on the left ankle, the spinous process of L4, the left wrist and the right wrist: for each, the resultant vector from the three axes is computed and then its root mean square. A high-speed camera at 100 frames per second filmed the movement, and everything was synchronised in the Nexus system. The release was defined as the instant the fingers on the string straighten. The analysis window runs from half a second before to a tenth of a second after the release, chosen on the basis of earlier work on string and arrow contact time (15 to 20 milliseconds) and on the windows used for electromyography and for the centre of pressure.

Strength was measured with a MicroFET 2 handheld dynamometer, always by the same examiner, in a fixed sequence: lower trapezius (right then left), shoulder horizontal adductor, horizontal abductor, upper trapezius. The lower trapezius is measured lying prone with the arm diagonally overhead, aligned with the fibres; the upper trapezius seated, with a depression force applied to the shoulder; the horizontal adductor and abductor supine with the shoulder at 90 degrees of abduction, in neutral rotation and the elbow at 90, with the sensor at the distal third of the forearm. Three trials per arm with thirty seconds of rest, and the mean of the three used in the analysis.

Differences between the two groups were analysed with independent-samples t tests and effect sizes with Cohen's d; associations with Pearson correlation; and predictors of score with multiple linear regressions using forced entry, adjusted for sex, age, weight and strength, after excluding variables strongly correlated with one another.

Six arrows per archer is few: it is the number on which both the mean score and all the stability measures rest, and in competition a recurve archer shoots seventy-two in the qualification round alone. The authors list the small sample and the absence of international-level athletes among their limitations.

The comparison between recurve and compound has a problem the paper does not flag. The two groups shot at different distances on different faces, and the mean score per arrow, 9.1 against 8.3, is presented as a between-group difference with a large effect size: it is a difference between two tasks, not between two abilities. The same applies in part to the stability measures, because the compound movement, with the let-off at full draw and a mechanical release aid, asks for different work. The compound archers were also lighter (65.8 against 73.5 kilograms) and had a reversed sex distribution (7 men and 11 women against 12 and 7): the differences do not reach the threshold for sex but do for weight, and strength is reported in kilograms force without normalisation for body mass.

The regressions deserve a close look. Each sets ten or thirteen predictors against 114 arrows for recurve and 108 for compound, and treats every arrow as an independent observation although the arrows come from 19 and 18 people: the clustering within each archer is not modelled. The power calculation is presented afterwards, on the number of arrows already collected. No coefficient of determination is reported, so how much of the variation in score those models explain is unknown, and the standardised coefficients, between 0.24 and 0.35 in absolute terms, describe moderate associations. Of the four surviving predictors, none has a confidence interval far from zero: the tightest reaches minus 0.001.

The centre of pressure comes from a single plate under the left foot: it is not the archer's centre of pressure but one foot's. Some angular velocities before the release have enormous standard deviations relative to their mean, such as the bow-side wrist in recurve (12.4 with a standard deviation of 26.5 degrees per second), a sign that a couple of archers moved far more than the rest. The time over which all this is measured is extremely short, and the tenth of a second after the release also contains the effect of the bow's recoil, which is a consequence of the shot as well as a condition of it: a higher wrist angular velocity after the arrow has gone may be the cause of a low score or its effect, and the design cannot separate the two.

Other factors the authors put outside their study: wind was not systematically controlled, and no psychological or fatigue variable was considered. Finally, the practical suggestions the paper offers, unstable-surface training for recurve and trunk stability work for compound, were not tested here: they are proposals built on the results, not measured interventions.

What this study adds is that the moment when stability matters is not the same for the two bows. In recurve, both predictors of score fall in the tenth of a second after the release, when the arrow has already gone and the bow recoils: the movement of the foot under the body's weight and the movement of the bow-side wrist. In compound they fall in the half second before, in the forward and backward sway and in the steadiness of the drawing wrist. Put more directly, in recurve what weighs is how the departure of the arrow is received, and in compound how the archer stands beforehand.

The explanation the authors offer is mechanical. A recurve held at full draw with the whole draw weight gives back a bigger push at the release: they measured a forward and backward sway in the tenth after release almost twice as large in recurve as in compound (5.4 against 3.2 millimetres), with higher angular velocities at the ankle and wrists. The compound bow, with its let-off and mechanical release, asks less afterwards and, in this sample, showed a more mobile centre of pressure during aiming instead.

On strength the finding is narrow: these recurve archers were stronger than the compound archers in shoulder horizontal abduction and adduction and in the bow-side lower trapezius, while the upper trapezius was the same in both groups. It is not a comparison between types of athlete that can be carried elsewhere, because the two groups were doing different tasks, and it is absolute strength, not scaled to body mass.

Two cautions, finally, against reading too much in. The compound archers' 9.1 points per arrow do not compare with the recurve archers' 8.3: the first group shot at 50 metres on an 80 centimetre face, the second at 70 metres on a 122 centimetre face. And the associations found are moderate, measured over six arrows per archer and on national-level athletes aged 15 to 24 whom the authors themselves describe as still developing, with results that do not match those obtained on elite archers.