Where arrows land on the target: Kolayis 2014

The study does not measure a performance or compare archers: it looks at where arrows end up on the target face and counts how they are distributed. Eighteen archers, seven women and eleven men, each shot two ends of six arrows, twelve in all, at 18 metres indoors. The faces were photographed and measured, and each arrow was described by two things: the zone of the target it landed in and its distance from the centre in centimetres.

The two hundred arrows left after exclusions are distributed unevenly across the four quadrants. The upper left quadrant took 59, the upper right 68, the lower left 37 and the lower right 36. Overall, 63.5 per cent of the arrows landed in the upper half of the target against 36.5 per cent in the lower half, and the difference is declared significant; between the left and right halves, 96 arrows against 104, there is no difference.

Within the gold, that is, the arrows scoring 10 and 9, the distribution across quadrants is also uneven: 35 arrows upper left, 33 upper right, 17 lower left and 12 lower right. In the red and blue zones, by contrast, the distribution across quadrants does not differ.

The mean distance from the centre, measured for each score, is practically identical in the four quadrants: the 10s sit between 0.83 and 1.13 centimetres from the centre, the 9s between 2.06 and 2.40, the 8s between 3.56 and 3.81, the 7s between 5.28 and 5.45, and the 6s and below between 7.47 and 8.47. None of these differences is significant.

The finding is therefore a single one: in this group the arrows went high more often than low, and the best arrows clustered in the upper part of the target. Why, the study does not investigate.

The eighteen participants are seven women, mean age 16.66 years with 5.88 years of practice, and eleven men, mean age 22.33 years with 4.11 years of practice. The standard deviation of the men's age is 12.95 years, so the male group puts teenagers and adults together.

The protocol is simple: two ends of six arrows at 18 metres in an indoor range, during practice. After each end the target face was photographed from three different angles with a digital camera. The images were calibrated and converted into metric units with the image processing tools of Matlab, and for each arrow its position and its distance from the centre were derived.

The target was divided into four equal quadrants on the horizontal and vertical axes: zone 1 upper left, zone 2 upper right, zone 3 lower left, zone 4 lower right. Arrows landing off the paper were excluded from the analysis. The arrows were then grouped by score, from 10 down to 6 and below, and by the colour of the zone hit: gold for the 10s and 9s, red for the 8s and 7s, blue for the 6s and below.

The analysis was run in SPSS 15: a one-way analysis of variance for the distances from the centre and a chi-squared test for the frequencies across zones.

The study is short and descriptive, and its main weakness is that it treats two hundred arrows as though they were independent of one another. They come instead from eighteen archers who shot twelve each, and one archer's arrows resemble each other: if two or three shooters were going high that day, they alone are enough to produce the result for the upper half of the target. No analysis allows for this clustering, and no result is ever reported archer by archer.

Information that would change the reading is also missing. The size of the target face is not stated, so the distances in centimetres cannot be set against the scoring rings; the bow type is not stated, whether recurve, compound or barebow; and whether all the archers were right-handed is not stated, which matters when left and right on the target are being discussed. Neither the condition of the equipment nor the point in the session is reported.

Eighteen archers times twelve arrows makes 216 shots, while the tables count 200: arrows off the paper were excluded and their number is not given. Excluding precisely the worst arrows shortens the tail of the distribution and can shift the counts by zone.

There are inconsistencies in the text as well. In the results, the difference between the upper and lower halves is called significant and accompanied by a p greater than 0.05. In the discussion, 54.32 per cent of the gold arrows are said to have landed in zone 1, while the table reports 59.32. The group of participants is mixed in age and experience and is never split.

Finally, the study explains nothing. It observes that arrows go higher more often than lower and stops there: there is no data on sight setting, technique, fatigue or equipment that would allow anyone to say where the tendency comes from.

The observation that remains is that in this session the arrows landed in the upper half of the target almost two times in three, and that the best arrows were concentrated up there. It is a group tendency over one afternoon of practice, not a feature of archery: eighteen archers, twelve arrows each, no control over bows, target faces or settings.

The part that may genuinely be of use to an archer is the method rather than the finding. Photographing the face after each end, calibrating the image and measuring each arrow's distance from the centre in centimetres is a finer way of reading one's own shooting than the score: the score says how much was hit, the position says where the error lies, and those are two different pieces of information. This study applies that method and proposes it as an assessment tool, and that is where its contribution lies.

What cannot be drawn from it is any guidance on what to do. The mean distance from the centre was the same in all four quadrants, so the study does not even say that those going high are missing by more; it only says they go high. And since the sight, the equipment and the point in the session were not recorded, it does not separate a change of setting from a fault in the shot. In the registry it counts as an example of measuring arrow position, not as evidence about a cause.