Postural consistency in archers: Stuart 1990

In this Loughborough study, published in 1990, nine skilled archers (eight men and one woman) each shot six arrows while an optical system recorded the position of the head, the drawing elbow and the bow. They were split into two groups by their official British handicap: 25 to 34 for the elite group, of national or international standard, and 35 to 45 for the other group, of good county standard. One archer's records were discarded because the recording of the first arrow could not be read, so the results cover eight archers.

At the loose the head came back to almost the same point sideways (standard deviation of 2 mm in both ends) and much less consistently along the line of aim (33 and 52 mm). The two groups did not differ in this repeatability, and the trend ran the opposite way to what one might expect: the elite looked more variable. For the distance between bow and elbow there was a difference, but again in the opposite direction: 1240 mm on average in the elite and 1079 in the other group, with arrow-to-arrow variation of 45 mm against 11, significantly larger in the elite. The only comparison favouring the elite, lower elbow variability in the first end, was not repeated in the second, and the authors give it no weight. The analyses of variance did not separate the groups, and the only correlation between level and consistency (r = -0.85 for sideways head position in the first end) disappeared in the second (r = -0.14). Within the elite group, variation in head-to-bow distance ranged from 7 mm in one archer to 161 mm in another.

Shooting took place indoors, 15 metres from a 60 cm FITA face, in two ends of three arrows, after the six sighters allowed in competition to set the sight; from then on the archers were asked not to change their point of aim, and they collected their own arrows as in competition. Three small reflective markers were attached to the olecranon of the drawing arm, to the end of the bow's pressure button and to the head, above and to the side of the aiming eye.

The Coda-3 scanner recorded the coordinates of the three points for 20 seconds at 30 samples per second, with a nominal precision of 0.1 mm; in a static check the standard deviation was 0.45 mm on the lateral axis and 0.08 mm on the other two. The loose was identified from the sudden jump in the movement trace, and posture at the loose was taken from the sample just before it. Each archer's data were referred to their own mean head position; consistency is the standard deviation within each end, and the groups were compared with correlations, variance comparisons and a three-factor analysis of variance (group, end, arrow).

The sample is tiny: eight archers analysed, six arrows each, and the paper does not say how many were in each group. The authors call it a preliminary study. All were already skilled archers, so the comparison is between good and very good, not between beginners and experts, and level comes from the season's handicap, not from the accuracy of the arrows shot in the laboratory: no postural measure is related to where the arrows landed. Shooting was indoors at 15 metres, far from outdoor competition distances.

There is also a problem that the authors raise without resolving. Using simple geometry they show that at 25 metres a standard deviation of only 2 mm, with bow and elbow moving in opposite directions, would send almost a third of arrows more than 13 cm from the centre; yet these archers varied by tens of millimetres. The authors rule out measurement artefact, but another reading remains open: that markers on the elbow and head do not capture the geometry that actually points the arrow. At 30 samples per second, moreover, the sample taken as posture at the loose may come up to 33 milliseconds before it.

Among archers who were already good, repeating the position of head, elbow and bow more precisely at the loose did not distinguish the best from the rest, and the two archers at the extremes, the most consistent and the most variable, were both in the elite group. This does not mean posture is unimportant: all of them had a well-drilled posture, and the head returned sideways to the same point within a couple of millimetres. It means that at that level differences in performance probably lie elsewhere as well.

For a coach the most useful information is individual. The traces show patterns that differ from one archer to another in the last moments: a small shift just before the loose in some, a sustained displacement in others, a gradual drift in others still, such as the elbow creeping forward towards the target. Differences of a few millimetres cannot be seen by eye or on ordinary video; the authors regard them as information for working with the individual archer, not as a test of who is better.