Caffeine and archery scores: Türker 2026 study

Twelve licensed Turkish university archers, ten men and two women, shot twice at 18 metres indoors seventy-two hours apart: the first time without caffeine, the second an hour after taking a 200 milligram capsule. Each session had three practice ends and then ten scoring ends of three arrows, thirty arrows in all, out of a maximum of 300 points.

The mean total went from 198.58 to 250.08 points, that is from 6.61 to 8.33 points per arrow, a difference of 51.50 points (t of 5.557, p below 0.001). The spread narrowed considerably: the standard deviation of the total falls from 36.09 to 11.96, and that of the per-arrow mean from 1.20 to 0.39. End by end, the differences are significant in seven cases out of ten (ends 2, 3, 4, 6, 8, 9 and 10) and not significant in ends 1, 5 and 7, where p lies between 0.055 and 0.098.

The repeated-measures analysis of variance confirms the effect of condition (F 30.88 on 1 and 11 degrees of freedom, p below 0.001, partial eta squared 0.737), finds no effect of end (p 0.054) and no interaction between condition and end (p 0.773): the gap between the two days is the same from the start to the finish of the thirty arrow series.

The authors close by saying that this is preliminary evidence rather than proof of a cause, and call for randomised placebo-controlled studies.

The archers were on average 22.92 years old with 6.50 years of practice, 175.83 centimetres tall and 74.58 kilograms in mass; ten shot with the right eye and right drawing arm, two with the left side. Musculoskeletal injuries, cardiovascular conditions, metabolic disease, recent surgery and adverse reactions to caffeine were all exclusions. The sample size calculation in G*Power called for fifteen participants; twelve were recruited, and the authors state this as a limitation.

The design is a one-group pre-post study, always in the same order. The first visit covered consent, anthropometric measurements (height on a Holtain stadiometer, mass on a Tanita BC-418 A) and hearing thresholds on an Interacoustics AC40 clinical audiometer, to confirm hearing between minus 10 and 25 decibels from 500 to 8000 hertz, since the protocol uses auditory signals to pace the shooting. Standardised breakfast at 9 am, shooting at noon, no caffeine.

Seventy-two hours later, the same breakfast at 9 am, a 200 milligram caffeine capsule at 11 am, a sixty minute absorption period, ten minutes of general warm-up plus sport-specific warm-up, and then the same shooting protocol. The capsule was a commercially available supplement and corresponds, the authors say, to the caffeine in two cups of coffee.

The shooting is at 18 metres indoors with the archers' own competition equipment: Hoyt recurve bows from 66 to 70 inches, draw weights between 34 and 42 pounds, Easton X10 arrows chosen to each archer's measurements. The target faces are the indoor three-spot five-ring faces used for recurve. Each end of three arrows had to be completed within two minutes, with short intervals between ends, in an indoor range with constant lighting and no external distractions. Scores were recorded end by end. The analysis, run in SPSS 27, uses paired-samples t-tests end by end and a repeated-measures analysis of variance as the main reference, with the Greenhouse-Geisser correction where needed. There was no randomisation, no counterbalancing and no placebo.

The authors list the limitations honestly, and they are the ones that matter: no placebo, no randomisation, no blinding, so learning, repetition and the expectation of an effect cannot be separated from the caffeine. Anyone shooting the same protocol twice does better the second time, and here the second time is always the caffeine time. Compliance with the instructions, twenty-four hours without caffeine and steady sleep and eating habits, was not objectively verified, and habitual caffeine intake was not measured at all, although it is the factor that most changes the response.

To that list belongs one difference between the two sessions that is not the caffeine. In the first, the archers shoot at noon after the 9 am breakfast, the anthropometric measurements and the audiometry; in the second, before shooting, they do ten minutes of general warm-up plus a sport-specific warm-up, neither of which is described for the first. An initial total of 198.58 out of 300 at 18 metres, that is 6.6 points per arrow, is very low for licensed university archers with six years of practice: it suggests the first session was not a clean baseline.

The dose, too, is less precise than it looks. It is fixed, 200 milligrams for everyone, and the authors convert it to about 3 milligrams per kilogram using the group's mean mass; at a mean of 74.58 kilograms the sum gives 2.68, and for the lightest and the heaviest participant the actual dose differs.

The figures contain two inconsistencies. The second and third tables report two F values for the same effect of condition, 22.01 and 30.88, and two effect sizes that do not match, eta squared 0.50 in one and partial eta squared 0.737 in the other, with the baseline mean printed as 6.61 in one and 6.62 in the other. The abstract, moreover, says caffeine was ingested in each of the two sessions, while the methods say none was given in the first. Finally the study measures nothing of the mechanisms it invokes: attention, reaction time and neuromuscular function were not assessed, as the authors acknowledge.

The raw result is large, more than fifty points over thirty arrows, and that is precisely why it should be read slowly: no dose of caffeine moves an archer from 6.6 to 8.3 points per arrow. A jump of that size, in a design where the second session is always the caffeine session and also carries an extra warm-up, says above all that the first session was not a reliable baseline.

What the record can offer someone who shoots is a way of reading their own indoor score. Taken one by one, the ends give seven significant differences out of ten, and the three that are not significant are the first and two in the middle: over thirty arrows the variation between ends is considerable, and with twelve archers a very large difference is needed before a single end comes out different. The absence of an interaction between condition and end says something else useful: the gap between the two days did not grow towards the finish, so neither session showed a loss of consistency over the thirty arrows.

On caffeine, this study on its own licenses nothing. The authors say so themselves, calling the result preliminary and asking for placebo-controlled trials, and they cite literature in which higher doses in traditional archery and in rifle shooting produced no effect. The registry gives no guidance on taking substances, and here there is none to extract in any case.