Bow tuning is the work of matching bow and arrow so that the arrow leaves cleanly and the shots repeat with consistency. It is also one of the areas of archery with the most recipes in circulation, often written with workshop precision: the right centre shot, the right button tension, the nocking point set to the millimetre, an eighth of a turn here and another there.
The measurements that exist are few, and almost all of them come from mathematical models and shooting machines, but on some points they agree. In a recurve model checked against photographs and a shooting machine, centre shot and button have no optimum value, because the adjustments offset one another and several combinations give good arrow flight; in the same model, the weight at the back of the arrow matters more for tuning than the weight of the point. Nine German national squad archers did not notice arrows that had been lengthened by five millimetres, and the idea that a set-up has to be built around the archer comes from a few Olympians in an interview, not from a measurement. Here I go through the main parameters one by one, from spine to the compound, and separate what the evidence supports from what remains range habit.
The archer's paradox: where tuning starts
When the string goes, the arrow does not leave straight. Under the push of the string the shaft bends, flexes around the riser and, according to the models, clears it without touching; it then keeps vibrating for a good stretch of its flight. This is the archer's paradox, and it is one of the few things in tuning that the registry rates as corroborated. High-speed footage of sixteen world-class archers shows the string pushing the nock 13 mm towards the body in just over five thousandths of a second, and the arrow leaving the button within 8 ms while the string only lets it go at 14.6 ms: by the time the tail reaches the riser, the arrow is no longer resting on it.
For the tail to pass cleanly, two independent models give the same condition: the shaft has to complete a full cycle of vibration while it is still on the string, and how fast it vibrates depends on how stiff it is, which is to say on its spine. Everything about tuning the arrow follows from this. With a mechanical release, on the other hand, the only experiment that compared the two releases on the same arrows found that the arrow flexed far less than off the fingers, which is one reason recurve and compound are tuned differently.
Arrow spine
Spine measures how stiff a shaft is. The number printed on it is a static deflection: the shaft is supported over 711 mm, or 28 inches, 880 g is hung from its centre and the sag is measured in thousandths of an inch. A 500 spine deflects 0.500 inches, a 380 deflects 0.380 inches, or 9.65 mm, and the lower the number, the stiffer the shaft.
The basic rule is that spine has to be matched to draw weight and arrow length: a more powerful bow needs a stiffer shaft, and on the same bow a longer arrow needs a stiffer shaft again. The rule holds in every model that has studied it, but nobody has measured how much an archer's group opens up with an unsuitable spine, and in the registry it is reported. A recurve model checked against expert archers and a shooting machine also gives the scale: moving to the next spine size takes about 11 N of draw weight to stay in tune, while a recurve can only be adjusted within about 10% of its nominal weight, so at a given draw length only one or two spines will tune well.
The starting point is the manufacturers' charts, which cross actual draw weight with arrow length. That the best archers follow them comes from an observation made in passing: in high-speed footage of sixteen national squad recurve archers at the London Olympic test event, arrow stiffness was in every case what the manufacturer's chart predicted, and fifteen of the sixteen shot the same shaft model. That tells us they use the charts, not that the charts point to the best shaft. On the second question there is a numerical model from 2025, not tested by shooting, which reproduces the charts overall but finds them too stiff at the extremes, for long arrows and low draw weights, by up to four sizes; it also notes that the charts move in steps of one inch, while in many rows half an inch of arrow already changes the recommended shaft.
Even so, the printed number is not enough to predict how the shaft will behave on the bow. Three models over twenty-five years say so, and in the registry the claim is transferred, because nobody has yet shot the same shafts and compared their static ranking with their ranking on the bow. The most concrete case concerns different shaft models: in the same calculation an X10 550, 4% stiffer than a ProTour 570, needs less draw weight to leave cleanly, 175.5 N against 180, because the two shafts have different profiles. That is why changing arrow model means re-tuning, even at the same nominal spine. Stiffness also shows up in vibration: in the measurements available, stiffer shafts vibrate at a higher frequency; between one spine and the next, though, there are only a few hertz, less than 3%, and nobody has shown that measuring it helps in choosing a shaft.
What coaches call dynamic spine, the behaviour of the shaft as built, with its point and its length, on its own bow, is read by shooting, with the bare shaft test: you shoot a few unfletched shafts alongside fletched arrows and look at where the bare ones land relative to the group. With a recurve shot right-handed off the fingers, bare shafts to the left of the group indicate a stiff arrow and bare shafts to the right a weak one; for a left-handed archer it is the mirror image. A shooting machine found the same direction on fifty-nine X10s with a right-handed bow, where the stiffer shafts went left.
| Adjustment | Change tested | Draw weight that offsets it |
|---|---|---|
| Spine | One size | About 11 N |
| Shaft length | 705 to 715 mm | 4.5 N |
| Button preload | 3.5 to 5 N | 3.5 N |
| Centre shot | 0 to +2 mm | 3 N |
| Fletching | 11 to 3 grains | 3 N |
| Point | 120 to 100 grains | 0.5 N |
The directions are the ones that are taught: a longer shaft, a heavier point or a more powerful bow make the arrow behave as weaker. The sizes, on the other hand, come from a model of a single bow, and not all of them match what is taken for granted on the range; I come back to this in the section on folklore. For a starting arrow length you can use the draw length calculator.
Reading arrow flight: fishtailing and porpoising
Some faults the arrow shows while it is in flight, and they are told apart by the plane in which the tail oscillates. Fishtailing is oscillation in the horizontal plane: the tail snakes left and right, and it points to spine, to clearance past the riser and to the release. Porpoising is oscillation in the vertical plane, with the tail rising and falling, and the first suspect is the nocking point, followed by fletching contact with the rest.
Coaches fix the vertical plane first and then the horizontal one, because as long as the nocking point is out it muddies the reading of everything else. How precise it needs to be is another matter. The only work that has calculated it, a recurve model checked with a shooting machine, finds a height at which the arrow leaves without rotating vertically, but around that height the optimum is a gentle one: a millimetre higher or lower changes the arrow's rotation by about 8 degrees per second, and moving to the next spine shifts the best height by two tenths of a millimetre. An error of a centimetre, by contrast, shows at once: with the arrow raised 10 mm on the string, the paper tear at 2 m went from under a millimetre to 40. In the registry, then, the idea that the nocking point has to be set to the millimetre is contradicted, with a serious limitation: the model holds the string exactly at the nocking point, while an archer holds it with three fingers.
Next to the nocking point there is the nock. In the trials of a Japanese shooting robot, with a training bow at 18 m, a vertical spread appeared that the robot alone did not explain, and the authors put it down mainly to the nock fitting loosely on the nocking point: the arrow sits in a slightly different place each time and changes its angle. It is a reasonable hypothesis, but the nocking point was never changed in any test and the authors themselves point to another possible cause, a draw that varied slightly without a clicker; that is why in the registry the idea that a loose nock changes the height of impact is unstudied. A nock that snaps onto the string with a light click remains a sensible rule.
Reading the flight has a limit too. Archers chase the arrow that flies clean on the assumption that the cleaner the flight, the tighter the group, but the only direct test does not bear this out: with a compound in a shooting machine at 35 m, halving the arrow's lateral wobble did not change the group, because the point, much heavier than the rest of the arrow, strayed less than three millimetres from a straight line while the tail swung. The author adds a condition, that the wobble does no harm as long as it repeats from shot to shot. In the registry the idea that an arrow that wobbles less groups tighter is contradicted, but by a single test, and nobody has repeated it with a recurve, fingers and seventy metres. Visibly bad flight is still a signal to correct; at that distance, though, an arrow wobbling twice as much did not group any worse.
| What you see | Plane | Likely cause | First move |
|---|---|---|---|
| Tail snaking left and right (fishtailing) | Horizontal | Spine not matched, clearance, release | Bare shaft, then shaft length or button |
| Tail rising and falling (porpoising) | Vertical | Nocking point out of place, fletching touching the rest | Move the nocking point in small steps and read again |
| Arrows landing crooked in no clear pattern | Both | Overlapping faults, often the release too | Close the vertical first, then read again |
The detailed diagnosis, with the variants that confuse people most and a step-by-step bare shaft protocol, is in the dedicated article: Fishtailing and porpoising: what arrow flight says.
Brace height
Brace height is the distance between the string at rest and the deepest point of the grip. You adjust it with twists in the string: more twists shorten the string and raise the brace height, fewer twists lower it.
Brace height governs a trade-off. With a lower brace the string stays with the arrow for longer and the arrow leaves a little faster; with a higher brace the arrow leaves the string sooner, and coaches regard the bow as more forgiving of imperfections in the release. Each manufacturer gives a range of values for the length of the bow, and the right point for an archer is found inside that range, changing the brace height a few millimetres at a time and comparing groups at competition distance. Here too, as with centre shot and button, you are looking for a range that works rather than an exact number.
Tiller
Tiller is the difference between two distances measured perpendicular to the string: the one between the string and the base of the upper limb, where the limb meets the riser, and the same measurement on the lower limb. In Olympic recurve it is usually kept positive, with the upper limb a little further from the string, because the bow hand pushes below the arrow while the fingers pull around the nocking point, so the two limbs do not work under the same conditions. The manufacturer gives the starting value, and many archers adjust it by feel at full draw, watching how the sight moves vertically while aiming.
Tiller is set with the limb bolts, and turning them also changes the draw weight and the arrow's behaviour in the vertical plane, which is why the nocking point should be checked again after every tiller adjustment. What is known about recurve tiller comes from coaching experience more than from measurement.
What is folklore and what is science
Many tuning protocols ask for a precision that the measurements do not support. The evidence is thin, almost all of it models and machines, and it leaves many questions open, but on some range rules it does say something specific.
The right value for centre shot and button
On a recurve archers look for the right centre shot, the lateral position of the point relative to the plane of the bow, and the right button tension, as if there were a number to find and any deviation would cost something on the group. The model behind the figures in Tab. 01, which studies the recurve in the lateral plane and was checked against photographs of the arrow and a shooting machine, looked for that number and did not find it. Moving the centre shot from zero to two millimetres requires going from 163 to 166 N of draw weight for the arrow to leave without rotation, and raising the button preload from 3.5 to 5 N requires going from 164.5 to 168: the adjustments offset one another, and over a reasonable range several combinations give good arrow flight. No combination cancels rotation and angle of attack at the same time, and the author advises tuning for zero rotation and accepting the small angle that remains.
The range has two edges. With a slightly negative centre shot the first, light contact of the arrow on the button disappears; a softer button absorbs release errors a little, but if it is too soft the arrow strikes it several times and the group opens up. In the registry the claim that there is no single right value for centre shot and button, only a range is transferred, because all the numbers come from calculation, and nobody has yet measured how wide that range is on the target.
The point as the main lever
To weaken or stiffen an arrow, archers often change the point. In the same model, however, going from a 120-grain to a 100-grain point moves the draw weight that tunes the arrow by half a newton, from 164.5 to 165, while going from 11-grain to 3-grain fletching moves it by three, to 161.5: six times as much, for less than half the mass removed. The author concludes that the point is not a useful tuning parameter and should be chosen for other reasons, such as wind or speed. The model does not explain where the difference comes from and nobody has checked it on the bench or on the target, so in the registry the claim that tail weight matters more for tuning than point weight is transferred. It does give one practical pointer: if changing the point moves the bare shafts less than you expect, shaft length and button are, in the model, much bigger levers.
The stiff side of the shaft and matched arrows
That every shaft has a stiffer side is documented: in 59 X10s of spine 325, flexed at every sixteenth of a turn, stiffness almost always follows a wave with two peaks and two troughs per revolution. The difference is small, though, 0.014 mm on average on a deflection of about 8.3, and in some shafts practically nil. Shot by a machine at 70 m, the shafts with the widest wave landed further right, while the position of the fletching relative to the stiff side moved nothing measurable; fletching to the label orients nothing anyway, because the label is printed at random relative to the stiff axis. In the registry stiffness that varies around the shaft and fletching position that does not move the impact are both reported, each on a single shaft model.
On the same arrows, how close they landed to one another was better explained by mass, mean spine and the frequency of that wave than by weight and nominal spine alone, and that is why in the registry weight and nominal spine are not enough to make a set uniform. Those same arrows, however, not matched by weight, stayed within a vertical spread of 8.8 cm at seventy metres, inside the 12.2 cm of the ten ring.
Small differences the archer does not feel
It is said that an experienced archer feels every slight change in equipment, and the feel of the shot is often used to choose and adjust it. An observation recorded in passing points the other way for small differences: in a study on the electromyography of the shot, one arrow in five had been lengthened by five millimetres, enough to move the click of the clicker, and none of the nine German national squad archers noticed. For large changes there is only a hint, from kyudo: with the string they normally used, archers mostly commented "feels as usual", while with strings of other materials the comments changed, but the study does not say they did not know which string they were using. In the registry the claim that archers notice large equipment changes but not small differences is reported, and neither observation measures a threshold. For tuning it suggests some caution: a minimal adjustment judged by feel should be checked on the group, given that top archers did not feel five millimetres of arrow.
The champion's set-up
That leaves the copied set-up. A 2020 study that interviewed eight members of the South Korean Olympic team reports the words of one of them: imposing the same shooting style and the same tuning on everyone does not help an archer improve, which is why this archer still turns to a former university coach for equipment. Another, shorter than the rest of the team and strongly affected by the wind, describes finding with the coach the string tension suited to their strength and a longer arrow. These are the opinions of elite archers in a single team, and nobody has compared archers with a set-up built on their own measurements against archers moved to a standard one: in the registry the idea that a set-up has to be built around the archer, not copied from a model is unstudied.
The few related measurements do not settle it. In eight competitive archers, shooting an imposed standard bow instead of their own did not change the precision of the group on average, and in the registry the claim that a different bow does not change precision is reported, on nine arrows per distance; in the individual cases, though, the weakest archer did worse with the heavier bow and the strongest did better. What can be said is that, if several combinations give good flight, a champion's set-up is one possible solution for that archer's bow, draw length and release, and there is no reason to expect it to be the solution for another archer.
The tuning sequence
With these limits in mind, the sequence used on the range still makes sense, as long as it is taken for what it is: an orderly way of finding a range that works, one adjustment at a time.
-
Base setup
Brace height and tiller within the manufacturer's guidance, a starting nocking point and centre shot, and a spine chosen from the chart for actual draw weight and arrow length.
-
The vertical plane
Set the nocking point by watching how the arrow leaves the bow, either with paper tuning, shooting through a sheet a few metres away and reading the tear, or with bare shafts.
-
Bare shaft
Comparing bare shafts with fletched arrows shows whether dynamic spine is stiff or weak, and you correct it with shaft length, button preload and, for small changes, point weight.
-
Walk-back
Shoot at increasing distances with the same aiming point: if the impacts drift sideways as you move back, adjust the centre shot.
-
Group test at competition distance
The final adjustments are judged on the group, at the main distance of the discipline, changing one thing at a time.
Whether a procedure like this raises scores has hardly been measured. In twenty-four Indonesian club archers, after sixteen sessions, the walk-back group went from an average of 443.9 to 547.0 points at 40 m and the paper tuning group from 419.6 to 435.1. The gap between the two methods looks large, but the authors did not test it statistically, no group shot the same sessions with the bow left as it was, and in the walk-back group the lowest score rose from 188 to 424 while the highest only went from 621 to 639: much of the gain may come from the archers who started lowest. In 2026 the same research group added a control group in a study of thirty compound archers, and the controls improved as well, from 213.7 to 236.2 points against a rise from 243.6 to 272.5 in the treated group, without the two groups being compared statistically. In the registry walk-back raising scores more than paper tuning and structured tuning raising scores are both reported: orderly tuning has been followed by higher scores, but it has not been shown to produce them.
When to re-tune
Once a tune is stable, you redo it from scratch when something in the system changes, and there are four such cases.
The first is replacing a major component: new arrows, new limbs, a new string. With a different shaft model the tune needs redoing even at the same nominal spine, because shafts with different profiles tune differently.
The second is a change in draw weight. In the recurve model one spine size is worth about 11 N, which is less than two and a half pounds, so even a few pounds more or less move the tune noticeably. After a jump in draw weight it pays to shoot the new weight regularly before touching the bow again, so that you do not compensate with equipment for an instability that comes only from adapting to it.
The third is a significant change in the release: if you have changed its dynamics, the system responds differently and the bare shafts need checking again.
The fourth is the seasons, for archers who compete in very different climates: intense cold and strong heat can change how materials behave, and checking the tune when the season changes is a coaches' precaution rather than a measured rule.
Compound tuning
A compound is tuned in a different way, because with a mechanical release the arrow flexes less and because cams, modules and cables add adjustments that a recurve does not have.
Cam timing is the synchronisation between the two cams. On two-cam bows you check it by seeing whether the draw stops make contact at the same instant at full draw, with a helper watching or on a draw board; when it is out, coaches associate it with groups that stretch vertically and a less clean arrow departure.
Cam lean is the sideways tilt of the cams under load, and it is corrected with twists on the two legs of the cable yoke, a job that takes experience and a bow press.
The peep has to line up with the eye at full draw. If it comes back rotated, you straighten it with a few twists of the string or by moving it between the strands, and you check it again whenever the string settles.
On the rest there is a model of the arrow's motion leaving a single-cam compound, checked against photographs rather than shooting tests. Moving the rest 2 mm sideways mainly changes the angle of attack, from 0.1 to -0.2 degrees, while the path of the arrow's centre of mass stays very similar, and the author concludes that rest position is not critical as long as the back of the arrow, and the fletching in particular, clears it. The margin depends on the cams: with the nock path of a binary cam the rest has to sit 4 mm below the nock instead of 2. In the same work the archer's torque on the bow matters much more, and in the registry the claim that rest position matters little on a compound, provided the arrow clears it is transferred.
Spine, too, seems to leave more margin with a release. In work by the same author, on a machine at 35 m, the recommended arrow and two stiffer sizes produced small, similar groups, all well inside a 40 mm circle, with only the sight setting changed: in the registry arrows stiffer than recommended grouping like the recommended one is reported, on one bow, one shaft model and arrows that were only ever stiffer, never weaker, than the recommendation.
Let-off is the share of force that drops away after peak draw weight. Measured in the laboratory on a production compound, it was about 65% at full draw, leaving roughly a third of peak force in the hand, and a model shows that changing the cams takes it from under 50% to about 80%: in the registry a compound letting off about two thirds of its force at full draw is reported, but on individual bows. The value is chosen on how the bow holds while aiming more than on an ideal number.
On the practical side, a compound needs dedicated tools, a bow press, a draw board, a bow scale and a level, that recurve archers usually do not own. Someone starting from scratch can invest in the tools or rely on a technician for periodic adjustments, because doing it yourself without a press, on a compound, causes more problems than it solves.
Whatever the bow, the measurements that exist point towards a habit good coaches already have: find a set-up that works within a range, keep it stable and know it well, instead of chasing an exact number that the measurements do not point to. Once the bow is stable, the work goes back to technique.