Where expert archers look: Kim 2019 study

Fourteen Korean archers, seven from the national team that had won Olympic gold in 2016 and seven university archers with no experience of national competition, watched a film of the shot recorded from the archer's point of view, while an eye tracker recorded where their gaze went. The film lasted 34 seconds and was split into six phases: standby, walking onto the line, nocking, waiting for the signal, raising and drawing, aiming.

In the first phase, before stepping onto the line, the national team archers made more fixations than the university archers (6.37 against 4.38) and fewer saccades, the jumps of the eyes from one point to another (8.00 against 11.47); they looked more at the target and the wind flags. From the second phase onwards the number of fixations did not differ between groups. The national team archers had longer fixations overall, and for everyone the longest fixations came during aiming, 5,666.57 milliseconds on average.

In a second experiment only the aiming phase was shown, for 8 seconds, with a beep or with a few coloured circles that appeared for a second around the sight. With both distractions the university archers made more fixations (3.25 against 2.47 with the circles, 3.04 against 2.54 with the beep) and more saccades (5.08 against 3.57 and 4.42 against 3.37), and their fixations were shorter (1,800.57 against 2,387.66 milliseconds with the circles, 1,857.60 against 2,342.06 with the beep). The type of distraction made no difference.

The film was recorded in the stadium of the national training centre with a Sony video camera and a GoPro 4, keeping the background and crowd noise; three coaches chose the most representative scenes. The archers, seated with chin and forehead on a rest, watched it on a laptop, and an EyeLink 1000 placed 45 centimetres away recorded the right eye only. A fixation was a period of at least 100 milliseconds with the gaze still within a window of 1.5 degrees; any shift outside the window counted as a saccade. For the first five phases the authors defined areas of interest: target, sight, wind flags, arrow rest and trees.

The first experiment had one familiarisation trial and three recorded trials; the second, two hours later, had 20 trials of 8 seconds, ten per type of distraction in random order. The beep came 2 to 4 seconds after the start, and there were 3 to 5 circles of different colours. Given the small sample, the data were converted to ranks before the analysis of variance, with non-parametric follow-up comparisons corrected by Holm's method.

The groups have seven people each, and some of the statistics do not fit this sample. The degrees of freedom reported (72 in the first experiment, 24 in the second) exceed what 14 archers allow, which suggests that trials or cells were treated as independent observations; the phase of the shot, which every archer watched in full, is also listed as a between-subjects factor. An interaction with F = 3.33 is reported at p < 0.001, a value those degrees of freedom do not give. In the second experiment the standard deviations are so small that they look like standard errors. In the discussion, one sentence credits the national team archers with shorter fixations in the early phases, while the rest of the text says the opposite.

The task is watching a video seated, head still, without a bow and without shooting, and the authors themselves note that on the range the gaze would move with the head and body. There is no measure of performance, so it is not known whether the national team archers' way of looking goes with better shooting. In the text all 14 are called elite-level archers, the university archers included. The discipline is not stated, but the Korean Olympic team shoots recurve.

The study describes two things archers will recognise. Before stepping onto the line, the Korean national team archers gathered more information, especially about the wind and the target, with a gaze that dwelt longer and jumped less; during aiming, when a distraction appeared, their gaze stayed steadier. The university archers were drawn away more, both by a sound and by something appearing near the sight.

This does not mean that training the gaze raises the score: the authors suggest training with distractions, but the study did not test it, and the task was watching a film. With seven people per group and statistics that do not add up, it is safer to read it as a description of how these archers looked during this simulation, useful for understanding what the strongest archers attend to in the different phases of the shot.