Arm movement and archery score: Taha 2016

Two right-handed Malaysian university recurve archers shot 18 arrows each at 70 metres with a small accelerometer strapped to each arm. From the acceleration the authors worked out how the arms moved and compared it with the score of each arrow, sorted into three groups: low (1 to 4), medium (5 to 7) and high (8 to 10).

The 36 arrows split into 8 low, 13 medium and 15 high, with four 10s. Aiming time got shorter as the score went up: 4.2 seconds on average for low arrows, 3.0 for medium and 2.9 for high, with standard deviations of 1.4, 1.4 and 0.4. At the release, the bow arm's maximum displacement along the forward and backward axis was 0.029 metres for low arrows and 0.014 for high ones, while the drawing arm went the other way, from 0.013 to 0.033.

Comparing the 10s and 3s of only one of the two archers, the authors describe broadly similar arm paths, with a larger movement at release and a steadier bow arm during aiming on the 10s. They conclude that a large horizontal movement of the bow arm at the release, and a longer aim, may lead to lower scores.

The accelerometers, built for the study, were fixed to both arms and sent data over Bluetooth at about 15 samples per second. Displacement was obtained by integrating the acceleration twice. The archers, at about 70 kg and 170 cm, with at least a year of university competition, shot three ends of six arrows at 70 metres, and the score and position on the target were noted for each arrow. The start of aiming was identified as the point where the arm signal settled after anchoring, and the release as the sudden movement on all three axes. For each score group the mean and standard deviation of the maximum displacement at release and of aiming time were calculated; no statistical test is reported.

The decisive limitation is the sampling rate. At 15 samples per second there are about 67 milliseconds between readings, while the release and the string's push on the arrow last around twenty milliseconds: the moment of release falls between two samples, and displacement values at that instant are not reliable. On top of that, integrating acceleration twice builds up errors very quickly.

The sample is two archers and 36 arrows, and the most detailed comparison, between the 10s and the 3s, covers a single archer. The authors repeatedly write "significantly" without reporting any test, and in the table the standard deviations are often larger than the means. Some values do not fit the reading given in the text: along another axis the bow arm moves 0.068 metres on low arrows, 0.011 on medium and 0.092 on high, which the text describes as very similar results. The categories also change between the summary (low from 1 to 3) and the body of the paper (1 to 4).

The explanations offered, such as bow arm fatigue or the bow tilting, were not measured.

For anyone who shoots, this study changes nothing. The idea that a steady bow arm and a not overly long aim go with better arrows is widespread among archers and coaches, but here it comes from two archers, with no statistics and an instrument too slow to see the release.

The only point of interest is technical: cheap sensors strapped to the arms can roughly tell apart the phases of the shot, such as anchoring, aiming and release. Measuring what happens at the instant of release, though, needs much higher sampling rates.