Balance and accuracy in compound archers: Acaroz 2026

Thirty compound archers of the Cyprus Archery Federation did two supervised balance sessions a week on a BOSU ball for twelve weeks, on top of their usual archery training, and were measured before and after on score, balance, grip strength and bow-arm shoulder joint position sense. The score over 15 arrows at 18 metres rose on average from 88.62 to 104.60 points out of 150, a mean change of 15.98 points with a standard deviation of 25.49: not everyone improved, and some got worse.

The study's question was not whether the BOSU work helps, but which changes went with the gains in score. After correction for multiple comparisons, four associations remain. Those who gained most in static balance on the right also gained most in score (Spearman's rho 0.594), and the same holds for static balance on the left (0.489) and dynamic balance on the left (0.514). The fourth concerns the bow-arm shoulder: those who most reduced their error in reproducing 125 degrees of flexion improved their score most (rho -0.505). In absolute terms, static balance went from 4.11 to 5.03 seconds on the right and from 3.49 to 4.50 on the left, and the position error in flexion at 125 degrees from 4.51 to 4.03 degrees.

Everything else is unrelated to the gains: dynamic balance on the right (0.379, which would have been significant before correction), the grip strength of each hand and of both together, the bilateral deficit, and joint position sense at 55 and 90 degrees in flexion and at any angle in abduction. Bilateral grip strength barely moved, from 44.93 to 45.96 kilograms. In an exploratory analysis, the baseline values of balance, proprioception and strength did not explain the baseline score either.

The study is a single-group longitudinal cohort with no control: the authors write that they could not set one up, given the limited number of professional archers available and their reluctance to withhold training from athletes in season. Data were collected at the outdoor facilities of the Cyprus Archery Federation between 10 August 2024 and 10 June 2025, with ethical approval from International Cyprus University and a ClinicalTrials.gov registration made after the study had started. The participants were 21 men and 9 women, mean age 28.2 years, twenty-four right-side dominant, nearly all with a right bow arm and a left drawing arm.

The programme ran two supervised sessions a week of 50 to 55 minutes, delivered by a physiotherapist in small groups: a static double-leg stance on the BOSU with one foot crossed over the other, single-leg stance, gliding stance, rising on the toes, and multidirectional reaching while standing on one leg. Static positions were held for 30 seconds, three repetitions, with 30 to 45 seconds of rest; dynamic tasks ran to 10 or 12 repetitions. Progression moved from two legs to one, increased the instability, brought in eyes-closed conditions and lengthened the holds. Attendance averaged about 85 per cent of the planned sessions, and anyone below that threshold was left out of the analysis.

The score is the total of five ends of three arrows at 18 metres, with the athlete's own compound bow and ten-ring scoring. Static balance is the Standing Stork Test, barefoot, hands on the iliac crests, heel lifted, three trials per leg with the longest kept; dynamic balance is the Y Balance Test, reaching anteriorly, posteromedially and posterolaterally, three trials per direction, normalised to leg length. Grip strength was measured with a digital Smedley dynamometer, seated with the elbow at 90 degrees, three maximal contractions of about 3 seconds per hand, then with both hands on the same instrument; the bilateral deficit is the difference between the sum of the two separate grips and the bilateral grip, as a percentage. Joint position sense was measured on the bow arm only, with eyes closed and a digital inclinometer: the examiner moves the shoulder passively to the target angle, holds it for 5 seconds, returns and repeats, and the athlete says when the angle feels reached; three repetitions at each of three angles, in flexion and in abduction.

The analysis is a Spearman correlation between change scores, with a Benjamini-Hochberg correction for multiple comparisons. The sample size was calculated in G*Power for an expected rho of 0.50 at 80 per cent power, which required 29 participants. The assessor did not deliver the sessions and did not see the change scores, but could not be blind to whether a measurement was the first or the second.

There is no comparison group, so the gains in score cannot be attributed to the BOSU work: everyone did the programme, and everyone carried on with their own archery training, which was not standardised. The authors say so plainly and call their findings hypothesis-generating rather than evidence that training balance raises scores.

There are thirty participants, and the change in score has a standard deviation larger than its mean: the spread between athletes is wide, and with numbers this small a few archers who improved a lot can carry a correlation on their own. Multivariable models adjusting for baseline score, sex and limb dominance were planned but not run, to avoid unstable estimates; the confounding from those variables therefore stays inside the results.

Balance was measured with field tests, barefoot on one leg, and not with a force platform during shooting: the authors themselves write that this may have missed the fine postural adjustments of the shot. The 125-degree angle is presented as a functionally informed approximation of bow-arm elevation, not as a measurement of the actual aiming position. Proprioception was measured twice in twelve weeks, with nothing in between. Psychological factors, vision, fatigue and shot-to-shot variability were not part of the study.

The score raises a question the paper does not address: 88.62 out of 150 over fifteen arrows is about 5.9 points per arrow, a very low figure for compound archers at 18 metres on ten-ring scoring, and the paper neither describes the target face used nor comments on the number. In the table of descriptive results, the change in abduction position error at 125 degrees is given as -0.84 degrees with a standard deviation of 0.12, far smaller than those of the two measurements it is derived from. The inclusion criteria speak of at least three years of practice while the results report four to eight. The version read here is an accepted manuscript that has not yet been through copy-editing, which the publisher flags as a possible source of errors.

The study describes which sensorimotor changes went with changes in score among national-level compound archers over three months. Those who improved most on single-leg balance tests, and those who most reduced their error in finding the bow-arm shoulder's high angle again, were also those who gained most points. That training balance raises the score is a different claim, and this design does not support it: nobody served as a comparison without the programme.

The negative half of the result is just as concrete. Grip strength, measured on a dynamometer with one hand and with both, moved by about a kilogram in twelve weeks and did not track the gains; the bilateral deficit, which measures how much is lost squeezing with both hands together compared with the sum of the two hands separately, fell without that having anything to do with the shooting. The authors suggest that in trained archers hand strength may already be past the threshold that matters, and that beyond it other things decide.

There is also a point about measurement itself: at baseline none of the tests predicted the score. In a homogeneous group of trained archers the single snapshot separated nobody, and only the change over time moved with performance. For anyone keeping records, that is guidance on how to measure rather than on what to train: the tests used here are a stopwatch on a single-leg stance, a three-direction reach and the task of finding a shoulder angle again with the eyes closed, all of them repeatable on a shooting field.