Posture and video feedback in archery: Hemara 2021

Thirty-one Thai recurve archers with international experience, on average 23.47 years old with 6.20 international competitions each, were filmed, reviewed their own shots with experts who corrected them individually, and were assessed before and after. The score over 10 arrows, out of 100, rose on average from 78.97 (standard deviation 6.04) to 85.61 (5.26), with t = 5.542 and p below 0.001. It is the study's only performance result.

The other figures concern an observation form built by the authors, the ASAAF, with 50 items grouped into ten 'skills' rated by observers on a scale from always to never; the mean score was 190.29 out of 250. Its correlation with the shooting score was r = 0.3753 before training and r = 0.4127 after, from which the authors conclude that the form explains 51% and 69% of the differences in scores. The paper also gives a ten-point technical description drawn from the videos, from foot position to follow-through, in which aiming lasts on average 4.60 seconds from settling on aim to the release.

The archers, chosen from those who had competed nationally and internationally, shot recurve at 70 metres. According to the procedure described, they first shot 3 ends of 5 arrows in an indoor sports hall, filmed by two cameras, and the videos were split into arrows that scored 10 and arrows that scored less. They then trained in the hall, watching the videos as often as they liked, with an expert correcting each of them individually until they felt confident; at that point they moved to the outdoor field of the Thai national stadium for the final test. The footage was analysed with Dartfish, and the camera listed is a consumer Full HD camcorder recording at 20 frames per second.

The ASAAF adapts a list of ten skills taken from another source: alongside concentration, coordination, balance and strength it includes items such as 'meeting like-minded people' and 'becoming a hunter and harvesting your own meat'. Scores before and after were compared with a paired t-test, and the relation with the form with correlation and regression.

The journal has questionable editorial standards, and the text shows it: much of the introduction covers the history of the bow, types of bowstring and tips taken from commercial websites, and the discussion is about the history of archery in Thailand. The design is a before and after with no control group, and the length of the training is not given. The first test appears to have been shot indoors and the second outdoors at the stadium, so under different conditions, and the ten-arrow score has low internal consistency (alpha 0.56 before and 0.66 after). The gain of almost seven points could come from the video, from the experts' corrections, from the extra practice or from familiarity with the test, and the study cannot separate them.

The statistics for the form do not hold together. With a single predictor the multiple correlation should equal the simple one, yet here it goes from 0.38 to 0.71 and from 0.41 to 0.83, and the 51% and 69% figures come from these values; moreover r = 0.38 in 31 people does not reach p < 0.001. The comparison between arrows scoring 10 and arrows scoring less, one of the three stated aims, has no results in the text, and the conclusion mentions augmented reality classes that the method does not describe. Finally, at 20 frames per second the release, which lasts a few hundredths of a second, cannot be analysed in any detail.

What remains for an archer is a general observation: archers who were already experienced scored a few more points in a ten-arrow test after a period of work with video and individual corrections from an expert. It is consistent with the idea that reviewing your shots with a coach helps, but with this design it does not demonstrate it, and it does not say how much the video counts compared with everything else.

The ten-point technical description covers a fairly standard recurve technique, and the 4.60 seconds of aiming is an average that is never related to the score. In the registry the study carries little weight, because of both its design and the journal.