Bodyweight circuit for archers: Yachsie 2022 study

Twelve male high-school archers from Banyumas Regency, in Central Java, in a single group measured before and after. Sixteen bodyweight circuit sessions between the first and the final test, with eight exercises: the regular, diamond, wide, pike, pseudo, archer, Hindu and bent-knee push-up. Two outcomes, the bow hold on the holding bow digit test and the score at fifty metres.

Both rise. The hold goes from 12.33 ± 0.88 to 14.08 ± 1.16; the score from about 219 ± 38 to about 284 ± 29, that is sixty-four points more. The article concludes that the bodyweight circuit improves endurance and accuracy and recommends it as work to be done at home.

The context explains the result better than the conclusion does, and it sits in the article's own introduction: during the pandemic the Banyumas archers "had never practised", and their monthly scores came out in the poor category, along with their arm endurance. A group resuming after months of a lay-off, measured once just after coming back and again sixteen sessions later, improves anyway. Without a group that resumes activity without the circuit, those sixty-four points cannot be attributed to the circuit.

Another thing does not add up, and it concerns the arithmetic. The table of tests gives t values of 34.019 for the hold and 41.894 for accuracy, on eleven degrees of freedom. From the published means and standard deviations the highest possible value, assuming a perfect correlation between before and after, is 21.7 for the hold and 24.8 for accuracy; for the hold, which is measured in whole numbers, the maximum falls to 13.4. To reach accuracy's 41.894 would need a before-and-after correlation of 1.024, which does not exist. The article's two biggest numbers, the ones its conclusion rests on, cannot come from the data in the table beside them.

A single group, measured before and after, with no control, and the article says so. Twelve athletes chosen purposively on four criteria: being elite archers, still being at secondary school, being male and being willing to take part. The contrast with the introduction is sharp, because there we read that these archers' monthly score falls in the poor category and that their arm endurance, measured with the hanging test from the Indonesian fitness battery, averages 8.3, also poor, with forty-five per cent of the athletes in the poor category.

The exercises are eight, and they are eight push-ups: regular, diamond, wide, pike, pseudo, archer, Hindu and bent-knee. One moves from station to station with no recovery until the circuit closes, then three minutes of rest. Ten repetitions, three or four sets, thirty seconds between stations. The load is set "using maximum heart rate, that is 220 minus age, then the intensity at 50-70 per cent": a percentage of maximum heart rate is not a way of dosing a circuit of push-ups, and in any case that range corresponds to a low intensity.

The sessions are eighteen, pretest and posttest included, so sixteen of training. Over how many weeks they are spread is not written, nor at what weekly frequency, so the length of the intervention remains unknown.

The instruments. The hold is measured with the holding bow digit test, with a declared validity of 0.961 and a reference to whoever proposed it: it is the only instrument in this batch carrying a validation figure, and that deserves acknowledgement. The protocol, however, is not reported, and the unit is not declared: the values sit between 11 and 16. Accuracy is the score when shooting at fifty metres; the arrow count is not written, and with a highest observed value of 307 it ought to be at least thirty-six.

The analysis is a paired t-test, with normality checked by Kolmogorov-Smirnov and a table titled "Levene Test Wilk Homogeneity", that is two different tests in one title. The threshold is written "p>0.05", with the inequality reversed. The critical value used, 1.80, is the one-tailed value on eleven degrees of freedom, while two-tailed it would be 2.20, and that the test is one-tailed is not stated.

The two t values cannot come from the data published beside them. The hold rises on average by 1.75, from 12.33 ± 0.88 to 14.08 ± 1.16. The t value of a paired test depends on the standard deviation of the differences, which is smaller the more the two moments correlate, and in the limiting case of a perfect correlation it is 0.28: from there the highest possible t is 21.7, against the published 34.019. There is more: that measure is in whole numbers, as its minima and maxima say, 11 and 14 before, 12 and 16 after. If the individual gains are whole numbers and their mean is 1.75, the most compact possible arrangement is nine athletes at plus two and three at plus one, which gives a standard deviation of 0.452 and a t of 13.4. The published value is two and a half times that.

For accuracy the calculation is similar and tighter. The mean gain is about sixty-four points, with standard deviations of 37.98 before and 29.01 after. For t to be 41.894 the standard deviation of the differences would have to be 5.32, and to get there the correlation between before and after would have to be 1.024. A correlation cannot exceed one. Even assuming it to be exactly one, the maximum t remains 24.8.

On the same page the text and the table swap the two numbers: the table assigns 34.019 to the hold and 41.894 to accuracy, the text says 41.894 concerns arm endurance and 34.019 concerns "arm muscle strength", which is not among the things measured. The abstract, in its turn, writes that the computed t is 0.00, which is the p value.

The descriptive table has accuracy's means shifted by two digits: minimum 123, maximum 267 and mean 2.19 before; minimum 200, maximum 307 and mean 2.84 after. A mean of 2.19 between 123 and 267 does not exist, and the text writes 2.197 where it means 219.7. The error is one of formatting and corrects itself at a glance, but it passes through the table, the text and the review.

A third number does not close. The discussion says that after the intervention the mean score at fifty metres sits between 312 and 322, while in the study the final mean is 284 and the highest score for a single athlete is 307. That figure comes from the regency's monthly observations rather than from the study's data, and the text does not distinguish it.

On the design, the limit is the one the introduction itself recounts. These archers had not trained for months because of the pandemic, their score was in the poor category and so was their arm endurance. The first test therefore falls at the lowest point of a lay-off, and the final test sixteen sessions later, once shooting had resumed. Any group improves under those conditions, and without a comparison that resumes activity without the circuit, the improvement cannot be attributed to the circuit. The word "elite" applied to the sample, moreover, does not square with the description of those same people in the introduction.

One observation remains on the content of the training. The eight exercises are eight variants of the same movement, the push, and the outcome to be improved is holding a bow at full draw, which is the opposite movement. That one of the variants is called the archer push-up does not make it a shooting movement: it is a one-armed push with the other arm out to the side. The transfer from eight pushes to an isometric pull needs an argument, and the article does not discuss it.

This article yields no evidence that a bodyweight circuit raises scores, for three reasons that hold together: a single group with no comparison, two t values the published data cannot produce, and a first test taken on archers who had not trained for months.

Something useful remains, though, and it is the setting the study came from. The question was concrete: during the pandemic the field was closed, the Banyumas coaches believed archery conditioning was done only by shooting arrows, and the athletes were at home. A circuit of eight bodyweight exercises, with no equipment and no space, is a sensible answer to that situation, and the article has the merit of writing the programme out in full: eight stations, ten repetitions, three or four rounds, thirty seconds between stations, three minutes between rounds. Anyone left without a field for a few weeks has a ready-made scheme here.

Two practical cautions on that scheme. The first: all eight stations are variants of the push-up, so it is a pushing programme. An archer needs the opposite, because the string arm pulls and the scapula closes: a useful circuit would be balanced with pull-ups, rows, band pulls and isometric holding work. The archer push-up, despite its name, is a one-armed pushing exercise.

The second: the way of dosing the load proposed in the article, a percentage of maximum heart rate computed as 220 minus age, does not work for a circuit of push-ups. In those exercises the load comes from the body's leverage and the number of repetitions, not from heart rate. Anyone using this scheme would do better to set it by the repetitions they can complete with good form, stopping a couple of repetitions short of failure.

The number worth remembering, finally, is not the gain in score: it is the starting picture, that in that group forty-five per cent of the archers were in the poor category for arm endurance. It is the reason why basic physical work, in a young group that does none, has room to work.