Stabilisers and lateral bow movement: Leonardi 1987

At the Sport Science Institute in Rome, in the physiology and biomechanics department, an archer from the Italian national team shot fifteen arrows for each of three configurations of the same bow, a 66-inch, 52-pound Hoyt TD4: with no stabilisation at all, with stabilisers and torque flight compensators using soft elastomers set loosely, and with the same stabilisers using hard elastomers set tight.

A single-axis accelerometer mounted on the rigid part of the bow recorded only the lateral component of the acceleration, from the moment the limbs begin their elastic return until the arrow leaves the string, that is, over the first thirty thousandths of a second. The bow's lateral displacement was derived from that signal.

The trace reads in three phases. Up to 22 thousandths of a second the arrow, bending, presses on the pressure button and pushes it back, with a force that grows to a maximum when the button reaches the end of its travel. Between 22 and 26 thousandths the button and the arrow return elastically towards the starting position and the arrow is pushed to the left by a force that falls to zero as the contact ends; in this part the curve does not return to zero, and the authors attribute what remains to the bow's torsion. Between 26 and 30 thousandths the arrow is no longer touching the riser but is still on the string, and the lateral acceleration depends mainly on the bow's internal oscillations set off by the peak force on the button.

At 26 thousandths of a second, as contact with the button ends, the bow's mean lateral displacement was 0.94 millimetres with no stabilisers, 0.84 with soft elastomers and 0.70 with hard ones. The standard deviations, though, run in a different order: 0.35, 0.20 and 0.25 millimetres. That is precisely why the authors prefer the soft set-up: not because it moves the bow less, but because it moves it the same way from one shot to the next, and they consider repeatability more important than the mean displacement.

The accelerometer was single-axis and measured only the lateral component, fixed to the rigid part of the bow. To mark the two moments that matter, the archer releasing the string and the arrow leaving the string, the clicker and a short length of the string itself were used as switches. Since the phenomenon lasts only thousandths of a second, sampling ran at 2500 hertz, with an HP 6940B multiprogrammer; means and standard deviations were calculated on an HP 9845C with software written by the authors.

The three configurations were: the bare bow; the bow with a 24-inch central rod carrying standard Hoyt masses, two side rods and a 10-inch upper front rod with the same masses, plus Spigarelli torque flight compensators with very soft elastic elements set fairly loosely; and the same system with rather hard elastic elements set tight.

Fifteen shots per configuration, all by the same athlete, chosen because he could repeat many trials with practically identical characteristics. The displacement reported in the table is the one at the last time step analysed, that is, at 26 thousandths of a second, with mean and standard deviation; the paper gives a significance threshold below 0.05 in all cases without naming the test used.

One archer, fifteen shots per configuration, one bow. The study describes the behaviour of a system, not of a population of archers, and with a single athlete the effect of the equipment cannot be separated from his hand, his release and his familiarity with that particular set-up. The order in which the three configurations were tested is not stated, nor whether the archer had time to adapt to each.

The measurement is indirect: the accelerometer gives an acceleration on one axis only, and the displacement is derived from it. Torsion and oscillation outside that axis do not enter the number, and the authors themselves note that in the second phase the curve does not return to zero precisely because of the bow's torsion.

The differences are small: 0.24 millimetres separate the worst configuration from the best, with standard deviations of the same order of magnitude. The paper states that significance is below 0.05 in all cases but does not say which comparison was made or with which test. There is also a questionable reading of the standard deviation, presented as a 34 per cent probability that a single shot's error falls within the mean plus or minus one standard deviation: 34 per cent is the share of one side of the distribution only.

No arrow was assessed on the target. There are no scores, no groups, and no link between the millimetre of displacement and where the arrow landed. That repeatability of the bow's movement is the most important property of a stabiliser, as the authors write in their conclusion, is a premise of theirs rather than a result measured here.

The equipment is nearly forty years old: today's risers, elastomers and stabiliser geometries are a different matter, and the absolute values do not transfer.

What the measurement shows well is the scale of the phenomenon. Over the twenty-six thousandths of a second in which the arrow is still touching the button, the bow moves sideways by less than a millimetre, in all three configurations, including the one with no stabilisers. Anyone who thinks stabilisation exists to stop the bow from moving while the arrow is still on it will find an order of magnitude here that differs from the usual picture.

The second point is the basis on which the authors choose. The hard elastomers give the lowest mean displacement, the soft ones the lowest scatter from shot to shot, and the authors prefer the soft: for them, in a stabilising system, it matters more that the bow does the same thing every time than that it does little. It is a reasonable basis and worth knowing, but in this study it is a stated premise rather than a result: nobody checked that less scatter on the accelerometer turns into tighter groups.

For an archer today, finally, the absolute numbers are worth little. A 1987 Hoyt TD4 with elastomer compensators is not current equipment, and what can be taken away is the method, that is, looking at the variability from shot to shot and not only at the mean value.