Anyone who takes up archery is told almost at once that you do not grip the bow. The hand rests on the grip, the palm pushes, the fingers stay soft and, when the arrow goes, the bow falls forward into the sling. Two more rules follow, both more technical: the hand goes in the same place with the same pressure every time, and turning it even slightly sends the arrow sideways. The bow hand can be measured, though, with force sensors in the grip, electrodes on the forearm and cameras recording thousands of frames a second.
The measurements do not back every rule equally. Three independent groups have found that expert archers do not grip, although none of the three links the open hand to where the arrow lands. In two small studies, steadier pressure goes with better shooting. The only study that tested whether the way the pressure is spread tells you where the arrow will go did not find it. On torque the mechanics are clear, while measurements on recurve archers are almost absent.
What the bow hand does
At full draw the bow pushes into the hand with the whole force of the draw, and the hand passes that push back to the riser through the grip. While the arrow is on the string the bow stays in balance, because the push of one hand and the pull of the other cancel out, and even a moment the hand applies to the grip is balanced by the string. The authors of the first study of the grip concluded that the stretch in which the hand can change the arrow's flight is the very short one between release and the arrow leaving the string, so they measured the pressure over the last 150 milliseconds before release, assuming it represented what follows.
Within that movement the hand does three things, each measured with different instruments: it pushes at a certain point on the grip, it closes to a greater or lesser degree, and it can rotate the bow about its vertical axis.
Where the push goes
The first study to put sensors between the hand and the bow came from an American laboratory in the early 1990s. The grip carried fifteen small force sensors: seven on the back, where the palm rests, six on the finger side and two on the thumb side. The archers were fifteen Olympic recurve shooters: eight from the US national squad, with FITA scores above 1,250, and seven less accomplished archers below 1,150. Each shot 72 arrows at 30 m over about two hours, and the pressure was recorded in three blocks of six arrows: start, middle and end.
The picture matches the rule. Almost all the load went through the palm, while the sensors on the thumb side and the finger side recorded much lower forces; on the back of the grip, the higher sensors and those towards the fingers carried more. The hand supported the bow without closing around it, and did so in the same way in both groups. The better archers tended to load higher up, with values about twice as large, but the spread within each group was so wide that the difference did not survive the statistics, and the authors put part of it down to their heavier bows. Fatigue changed little either: after 72 arrows, at a faster pace than in competition chosen to bring it out, the authors found no systematic change in how the force was spread, and only the moment about the vertical axis rose slightly towards the end.
The muscles that open and close the fingers
A force sensor shows where the hand pushes, not whether the fingers close. For that you need the muscles, and a German and Turkish group measured them with electromyography on the bow hand's flexor digitorum superficialis, which closes the fingers, and on the extensor, which opens them. Ten recurve archers from the German and Turkish national squads and ten club beginners were recorded over twelve shots each, in the two seconds around the clicker.
In the experts the extensor worked harder than the flexor over nearly the whole window: the fingers stayed open because the muscle that opens them was active, not because the hand was switched off. There is one exception the coaching rule leaves out: between 100 and 300 milliseconds after the click the extensor's advantage disappeared, because both muscles peaked at around 200 milliseconds and began to relax straight after, so even in a national squad archer's hand the two muscles fire together for a moment just after the click. In the beginners, by contrast, the two muscles worked at the same level before the click, and between 100 and 200 milliseconds after it the flexor jumped sharply. The difference between the groups in the flexor only became clear at 500, 600 and 700 milliseconds after the click, when the arrow had already left the bow.
A third group, thirty-two young Malaysian archers split into higher and lower scorers, found the finger extensor of the left arm more active in the better group, with a p value of 0.050, exactly on the threshold; the study does not say, however, which was the bow hand. The three studies reach the same conclusion from different sides, which is why the registry rates the claim that expert archers do not grip the bow as reported: these are comparisons between groups of people, and none of the three sets the bow-hand muscles beside where each individual arrow lands. One arrow-by-arrow comparison does exist, and it complicates the picture.
The hand muscles, arrow by arrow
A French group followed eight high-level recurve archers through 549 arrows at 70 m, with electromyography on many muscles, and sorted the arrows into three groups: 8 or less, 9 and 10. In the bow hand the finger extensor was more active on arrows scoring 8 or less than on those scoring 9 and 10, both while aiming and in the follow-through. The flexor digitorum superficialis did the opposite, and was less active on the worse arrows than on the 10s.
The authors do not read this as a more relaxed hand on the better arrows. On the low arrows the extensor showed sporadic bursts of activation that were absent on the high ones, and a flexor that does too little, in their view, costs control: for both muscles they speak of the right level of activation, and they say so cautiously, because the result does not match the comparison between elite archers and beginners. A related hint comes from four traditional archers, tested in a Malaysian laboratory, who tried four different bow grips: in all of them the grip with the best result was one in which the extensor worked harder than the flexor, but the best result did not necessarily go with the highest muscle activity.
That is why the claim that on better arrows the bow hand is more relaxed is contested: the same study points two ways, one for each muscle. The two findings sit together better than it might seem, because keeping the fingers open while the bow presses into the palm takes an active muscle, and on the French archers' worse arrows what made the difference was sudden bursts from that muscle.
How the pressure is spread and how it repeats
The main question of the grip study, though, was a different one: whether the way the hand spreads its pressure predicts where the arrow goes. The authors had fixed the expected direction in advance, with more load at the top supposed to lower the arrow and more load on the finger side supposed to move it sideways. For each archer they compared the eighteen recorded arrows with where they landed. The expected link appeared in one archer out of fifteen for up and down, and in three out of fifteen for left and right; in the rest the correlations were small and did not even always point the expected way.
So the registry writes the claim in the direction of the result: how hand pressure is spread over the grip does not predict where the arrow goes. Its status is reported and no more, because there is a single study, with limits the authors list themselves: fifteen sensors, seven of them on the back, may miss subtle changes; the arrow's deviation was measured from the centre of the target rather than from where the archer was aiming; and the release, which moves the arrow in its own right, was not controlled. Eighteen arrows each also leave room for a small effect. The authors keep two readings open: either the instrument was too coarse, or the normal variations in an expert's hand are too small to move the arrow.
What the same study did find concerns consistency. The average distribution of pressure did not separate the two groups, but its variability from arrow to arrow did: in the national squad group the point at which the hand loaded the bow vertically changed less from shot to shot, with a statistically significant difference, while horizontally the groups were alike. And it was vertically that the better archers landed closer to the centre, 2.7 centimetres against 5.7, a coincidence the authors note without turning it into a chain of causes.
The same picture returns in kyudo. Nineteen Japanese university practitioners shot eight arrows at 28 m with an instrumented bow that recorded the force on the grip, the upward push and the torsion: those who hit most often applied force and push that repeated closely from shot to shot, while those who hit least varied a great deal. The grip and the hold differ from the recurve's, but the direction is the same, which is why the claim that better archers apply a more consistent pressure on the grip is reported: two associations on small samples, and nobody has yet tried making the pressure steadier to see whether the group tightens.
Bow torque
If the hand pushes more on one side, or the wrist moves off the plane of the bow, the grip receives a twist about its vertical axis and the arrow leaves at a different angle. The model figures are striking. A model of a 166 N recurve bow calculates that thirteen hundredths of a degree of difference from shot to shot moves the arrow by more than one ring at 70 m, and corresponds to about 1.7 millimetres of misalignment between the arrow point and the front stabiliser, seen at full draw. On a mechanical launcher with a Japanese bow, varying the torque alone, up to about 14 newton metres, moved the point of impact at 28 m from one side of the target to the other, the test behind the registry's claim that in kyudo the twist the hand gives the grip shifts where the arrow lands.
In all these tests the archer is missing. The only direct measurement of torque on people comes from the instrumented kyudo bow: the torque is there, it is worth about six tenths of a newton metre, less than a twentieth of the torque used on the launcher, it grows with the force on the bow and, in the last five hundredths of a second before the arrow leaves, it behaves differently depending on how often the archer hits. Nobody has measured torque and point of impact on the same shots of a recurve archer, so the claim that twisting the bow with the hand moves the arrow on the target is reported, and it sits among the equipment claims because much of the evidence comes from models and test benches.
How common torque is at the top is the subject of a single study. At the test event for the London Games, sixteen recurve archers from national teams, eight men and eight women, were filmed at 4,000 frames a second. In none of the sixteen did the bow move sideways by more than a millimetre, or rotate by more than a tenth of a degree, before the arrow left the string, a result the registry keeps separately: in high-level archers the bow does not move measurably while the arrow is on the string. The authors therefore inferred the torque from the arrow, looking at the ratio between its first and second flex and at the distance by which the back of the arrow cleared the button.
In nine archers the ratio was between 0.6 and 0.8 and the back of the arrow cleared the button by 8 to 10 millimetres, values consistent with a bow that was not twisted. In five the ratio fell to 0.4 to 0.6 and the clearance to 2 to 5 millimetres, the likely sign of an anticlockwise twist seen from above; in two it rose to 0.9 and 1.0, with a clearance of 12 and 25 millimetres, the sign of a clockwise twist. The footage often showed where it came from: in the more marked clockwise case the wrist sat clearly to the left of the plane of the bow, as seen by the archer, and in the other a thumb out of line with the target pointed the same way; in three of the five anticlockwise cases the tip of the thumb pressed on the side of the riser.
The claim that even among elite archers, many show signs of twisting the bow is therefore reported, with two clear limits: the twist is the authors' subjective judgement, made by looking at the arrow and the hand, and the study does not say what it costs in points. The authors even add that such a fault, repeated identically on every arrow, is compatible with impressive scores.
What I take to the shooting line
These studies do not hand over a shape of hand to copy. They say that in expert archers the push goes through the palm and the fingers do not close, that consistent pressure goes with better shooting more than the way the pressure is spread, and that on paper a small twist is enough to move the arrow. None of these results comes from an experiment that changed an archer's hand and then measured the score, and I keep that in mind in the three points I draw from them.
The first concerns where the hand sits. Since no study has found one distribution of pressure better than the others, there is little point in chasing an ideal hand position; it makes more sense to choose one in which the palm holds the bow without the fingers having to help, and to work on bringing it to the same place on every arrow. A mark on the grip, or a photo of the hand at full draw taken from the same angle every time, helps to show whether that place drifts during a session. It is part of the repeatability of the shot.
The second concerns the fingers, and here the measurements call for caution. In experts they stay open because the extensor is working, while in beginners the flexor jumps after the click; in the arrow-by-arrow comparison, though, the flexor was more active on the 10s than on arrows scoring 8 or less, follow-through included, and the worse arrows had sporadic bursts from the extensor. Keeping the fingers extended without forcing them therefore looks closer to what experts do than opening them hard, and nobody has shown that a hand closing after the arrow has gone costs points. A slow-motion video shows it clearly, but on the data we have it is not the first problem to solve.
The third concerns torque, and it is the easiest to check without instruments. In the London footage the two visible causes were the wrist off the plane of the bow and the tip of the thumb on the side of the riser, and both show up in a photo taken from behind or above at full draw. The models also say that what moves the arrow is torque that changes from shot to shot, not torque that stays the same, and the seven archers showing signs of a twist were still members of national teams. So before changing the shape of the grip, I would look at whether the torque changes from one arrow to the next.
What is missing
The grip study needs repeating with modern sensors, more arrows per archer and competition distances, recording where every arrow lands and separating the two questions: whether what matters is how the pressure is spread, or only that it repeats. An instrumented riser is needed to measure torque on the same shots in which we see where the arrow goes, and so is an experiment with a control group that trains a consistent hand and looks at the score after a few weeks. On the muscles of the hand, finally, the comparisons between people and the one between the same archer's arrows do not agree, and a replication from another laboratory should measure grip pressure and electromyography together, because a relaxed hand and a hand that always pushes the same way are not the same thing.