Final draw, clicker and reaction time: what research has measured since 1993

The question

In the last two or three seconds of the shot a recurve archer is doing two things at once: holding the aim and drawing the arrow point a few millimetres further, through the clicker. Research has looked at this final phase from two angles. The first is the final draw itself: how smooth the movement is, and whether that smoothness shows up in the score. The second is the time between the clicker snap and the release, and whether a quicker reaction produces better arrows.

This page follows the studies in chronological order, with their numbers and their limitations. Practical advice is in the articles on the final draw, the clicker and the release; this page sets out where our current knowledge comes from and how firm it is.

1993-2006: the final draw becomes measurable

The first study came from the physiology and biomechanics laboratory at INSEP in Paris. Leroyer and colleagues fixed a transducer to the bow that recorded, indirectly, how far the draw hand moved, and had five elite and three university-level archers shoot 36 arrows each at 90 m. Over the last five seconds the hand came back by about 5 mm, slowing to roughly a millimetre per second. What separated the archers was how smooth that movement was: less smooth in the university archers than in the elite group and, in every one of the eight, less smooth on the worst arrows than on the best.

The same paper reported two smaller experiments. In the first, seven archers held at full draw until exhaustion, at around 40-50% of their maximal force: a peak between 8 and 12 Hz, the frequency of physiological tremor, appeared both in the electrical signal of the posterior deltoid and in the movement of the bow. In the second, a tone sounded whenever the draw hand gave way beyond a set threshold: smoothness improved in five archers out of six, but the score did not improve significantly.

In 2006 Jürgen Edelmann-Nusser's group noted that Leroyer's was still the only study of draw length during the final draw, perhaps because his equipment was slow and awkward to fit. Their own solution was simpler: a strain gauge glued to the clicker, which bends more or less depending on where the arrow point sits beneath it. Seven German juniors who had competed at the Junior World Championships and the German national championships shot 66 arrows each at 30 m indoors. The archer with the highest average drew the arrow through the clicker in one continuous movement over the last two or three seconds; the two with the lowest averages showed, on precisely their 5s and 6s, a loss of draw length just before the click (what the authors call shrinking) or a pause. Using an optical system mounted on the riser, the same authors also tracked the aim: for every archer, movement in the final second was smaller on 10s than on 8s or lower, but across archers the average amplitude did not tell you who scored more.

Both 2006 papers describe the same seven archers, so the registry counts them as one data set. The final-draw part is also a reading of the curves with no statistical test: it points the same way as Leroyer, without measuring the effect.

The clicker: a signal to react to?

Leroyer described the clicker snap as the stimulus the archer responds to, and that was the accepted view. The interval between click and release had already been measured in electromyography studies, with a sensor on the bow marking the release: 170-180 ms in the most skilled archers of a group of twelve studied by Nishizono in Japan, and 185 ms on average in the fifteen archers, most of them on the US team, studied by Martin and colleagues.

In 1998 Edelmann-Nusser and Gollhofer, in Stuttgart, challenged that view. Nine archers from the German national squads shot 35 normal arrows and five "extended" shots, in which they had to wait more than a second after the click before releasing. The EMG pattern of eight muscles was the same in both cases. The click-to-release interval was also shorter than a simple reaction time to a sound in seven archers out of nine, and shorter in all nine than the reaction time measured while they held a laser dot steady on a target. The authors concluded that the motor programme for the release starts before the click, and that the clicker works as a check on draw length and a confirmation of the shot just taken rather than as a starting signal. It is a short conference paper, with analyses run archer by archer and few numbers reported.

A clue pointing the same way came twenty-five years later, from the feet. In the study by Kuch and colleagues, six elite and seven sub-elite archers shot 18 arrows at 70 m standing on two force plates. In two shots out of three the centre of pressure started moving to counter the bow's recoil before the release, and on those shots the elite archers started on average 23 ms after the click: shorter than the latency of a postural reflex, which the authors put at around 40 ms. In their reading, the command had already been programmed.

In 2023 and 2024 a group at Seoul National University took the question into the laboratory, with eight men who had never shot, using a bow frame fitted with sensors. A self-timed release produced tighter virtual shots than a release triggered by an external cue (a precision index of 0.017 against 0.023), and the second paper links precision to coordination between the hands and between the fingers. The two papers report the same participant characteristics and count as a single data set; and since nobody shot a real arrow, the comparison with the clicker remains an analogy offered by the authors.

Clicker reaction time and score

A measurement designed for the field arrived in 2005, when Hayri Ertan's group at the Middle East Technical University in Ankara validated a chronometer that fits on the bow, which the same group had developed in 1996: a switch under the clicker starts it, and a sensor on the bow stops it when the arrow point passes in front of it. In 20 elite Turkish archers the chronometer gave a mean of 177 ms and finger-extensor EMG 146 ms, and the two measures moved together (r = 0.79). The gap is not an error: EMG marks the start of muscle activity, while the chronometer marks a moment after the release, when the arrow is already moving. Two years earlier, using EMG, the same group had compared ten elite archers, ten beginners and ten people who had never shot: the finger extensor peaked about 100 ms after the click in the elite group, 200 ms in the beginners and 300 ms in the non-archers.

In 2010 and 2011 two papers with Cevdet Tinazci among the authors combined the chronometer, EMG, force plates and an aiming system on four elite archers at 18 m, analysing 108 arrows scoring 8, 9 and 10. They share the same tables, but the first describes four women and the second four men. According to the tables, on 10s the finger flexors were less active in the last second before the click (80% of maximal contraction against 85.6% on 8s) and the draw was shorter (4.8 against 5.6 s), while clicker reaction time was the same for 10s, 9s and 8s, at around 142-143 ms. The abstracts, however, state that reaction time is shorter on better shots, and the second paper has been cited that way, including by the Australian study of 2017. The registry goes by the table: no difference at group level, and a faster reaction linked to better arrows in only two of the four archers taken individually.

That Australian study is also the largest. Before a World Cup event Spratford and Campbell, of the University of Canberra and the Australian Institute of Sport, tested 39 elite recurve archers from four countries, ten arrows each at 70 m, with a force plate, high-speed cameras and a microphone to pick up the click. Mean clicker reaction time was 163 ms and, together with bow draw force and the speed of body sway after the release, it explained 42% of the variation in arrow score. The authors themselves play down its weight: reaction time varied little between archers (a standard deviation of 17 ms) and its effect spanned only about half a point. Below 102 ms, the lowest value recorded in the 2005 study, they suggest it would already be anticipation.

The later studies at 70 m mostly point the same way. In Kuch's study the elite group reacted to the clicker faster than the sub-elite group (146 against 156 ms). In 2025 Jacquot and colleagues tested eight top French archers in competition-like conditions, 72 arrows each outdoors with their own equipment: across the 549 arrows analysed, reaction time was longer on 8s or lower than on 10s, while archer by archer the difference was not significant. In Jacquot's doctoral thesis, defended at the University of Lorraine in 2026, the same eight archers were compared with seven archers from the regional circuit, again at 70 m, and clicker reaction time did not differ between the two groups.

Indoors, the same interval has been looked at from another angle. In 2016 Callaway and colleagues filmed fifteen archers of mixed standard, from university to international level, for 30 shots each at 18 m, with a microphone to pick up the click. Every variable in their model was expressed as the shot's deviation from that archer's own average, and the deviation in click-to-release time was the first to enter; the whole model, however, explained only 7.7% of the variation in the arrows' distance from the centre.

The Ankara group came back to the clicker in 2021 with EEG. Across 15 recurve archers and 1,070 arrows at 18 m, the click produced the classic auditory response in the brain (the N1 component), larger on arrows that hit the centre; the timing of the response did not differ between good and bad arrows. The ethics approval carries a 2004 reference number, and the differences are judged with 95% confidence intervals for two zones of the target and 90% for two others, with no correction for multiple comparisons.

Only one study in this line has tried to change reaction time through training. In Turkey twenty boys aged 15-17 were split into two groups, and the one that did eight weeks of strength training went from 93.5 to 105.5 points over 12 arrows at 18 m, with shorter simple reaction times; the control group went from 94.1 to 95.4. But the reaction measured was to a light and a sound, not to the clicker, the paper does not say that the groups were randomised, and the two groups are never compared with each other. Outside clicker-equipped recurve there is a study of ten Turkish traditional archers: stopping at 98% or 96% of full draw before the final draw gave the fewest micro-jerks in the hands and bow while aiming, but the shots were taken from 1.5 m and not scored.

What holds up today

Within the same archer, a smooth final draw goes with better arrows: Leroyer saw it in all eight of his archers, and Edelmann-Nusser's reading of the curves points the same way, with the clearest irregularities appearing on the worst arrows of the two weakest juniors. These are two independent groups that agree, but they are small, both predate 2007, and the second used no statistics.

On clicker reaction time, three independent studies of elite archers at 70 m (2017, 2023, 2025) find shorter times on better arrows or in better archers, always at group level. The comparison between archers of different standards is not consistent, though: Kuch found the elite group faster, while Jacquot's thesis found no difference between elite and regional archers. Where they are reported, the differences are around ten milliseconds, and the estimated effect on score is about half a point. On individual arrows, the only contrary data we have read, from Erkek and Tinazci, come from four archers with each arrow treated as an independent observation; under the registry's rules, one weak contrary study among supporting ones does not make a claim contested. At 18 m, in Callaway's study, what counted was how far each shot's timing strayed from the archer's usual value, and that too carried little weight.

The values measured fall almost entirely between 140 and 185 ms: 142-143 ms in Erkek and Tinazci's data, 146-156 in Kuch's, 163 in Spratford's, 170-180 in Nishizono's, 177 with the 2005 chronometer and 185 in Martin's. They can only be compared roughly, because each study marks the release in a different way.

What remains open

Nobody has trained clicker reaction time to see whether the score changes: all the supporting data are associations, and the one intervention measured a different kind of reaction. The direction of the link is also open. If the release is prepared before the click, as the 1998 study, Kuch's force plates and the Seoul bow frame all suggest, a short time may be the sign of a shot that is already well under way rather than of a quick response.

The measurement is not standardised. The moment of release has been marked with a sensor that detects the arrow point going past, an inertial switch or an accelerometer on the bow, a strain gauge, a microphone, a high-speed camera, video analysed by hand or the activity of the finger extensors, and each falls at a different point in the movement: within a single study, EMG and chronometer differ by thirty milliseconds, more than the gap between elite and sub-elite archers that Kuch found.

Smoothness of the final draw has been measured by only two groups, on fifteen archers in total, and the one attempt to train it, Leroyer's audio signal, made the draw smoother without a significant improvement in score. Almost all the studies involve clicker-equipped recurve bows: there is nothing on compound or barebow, and for traditional bows there is only the Turkish study at 1.5 m. There is only one direct comparison of shooting with and without the clicker in archers: in nine world-class Korean archers, shooting without the clicker made the draw shorter and the aim more consistent, but at 8 metres, over twenty arrows per condition and with no score at a competition distance. On the link between a quick reaction and score, finally, studies in the registry cite the 1996 paper in which the Ankara group developed the chronometer in opposite ways, some as evidence of the link and others as a study that did not find it, so the line rests on the measurements each study reports for itself.

@misc{campagna2026_enresearchfinaldrawclicker,
  author = {Campagna, Matteo},
  title = {Final draw, clicker and reaction time since 1993},
  year = {2026},
  howpublished = {Research, matteocampagna.com},
  url = {https://matteocampagna.com/en/research/final-draw-clicker},
  note = {accessed 2026-10-16}
}