"Get the elbow behind the arrow." It is the first correction an archer hears, usually in the first few weeks, and it stays with them for years: the elbow of the drawing arm should come round behind, until it sits on the line that runs from the bow hand through the string hand, and it should not rise too far above the shoulder. The mechanics behind the rule are simple: if bow, hand and elbow lie on a straight line, the force of the string runs along it, the back holds it and the arm does not have to correct sideways.
The question has two parts. The first is descriptive: where the elbow of strong archers sits when it is measured, and how much it varies between archers. The second is the one that matters on the range: whether bringing it closer to the line of force earns more points. On the first, the studies give some precise figures. On the second, the registry currently marks the claim as contested, and the reason lies in how the studies were done more than in how many there are.
What is measured when people talk about the line of force
Seen from above, an archer at full draw forms a narrow triangle: the bow hand in front, the string hand at anchor, the elbow of the drawing arm behind. The line of force is the straight line from the bow hand to the string hand. The elbow is aligned when the forearm of the drawing arm continues that line, and it deviates when it moves out to the side; the deviation is measured as an angle, in degrees.
The studies, however, do not all measure it the same way. In a Korean squad the angle is the one between the arrow and the line running from the bow hand to the wrist and elbow of the drawing arm; in a group of young Malaysian archers it is the angle between two segments read off a video shot from the side, from the bow wrist to the string wrist and from the string wrist to the point of the elbow; other studies measure the elbow angle or how far the shoulder is opened. A single experienced archer filmed with two cameras, from above and from the side, shows how much the viewpoint matters: from above the angle between elbow and shoulder was 45 degrees, with the elbow off the line, while from the side the line was almost straight. Two studies looking from different views are not measuring the same thing, so each figure below goes with its own definition.
Where the elbow of strong archers sits
In the archers measured, the elbow hardly ever sits exactly on the line. In 31 Korean elite recurve archers at full draw, the line of force deviated from the arrow by about ten degrees: a mean of 10.06 degrees in one group and 9.20 in the other, before any intervention. In 16 young recurve archers from a Malaysian state squad, shooting at 70 m, the elbow deviation at anchor averaged 4.26 degrees, with arrows ranging from 1 to 8. In these squads a small deviation is the norm, not the exception.
One figure in the Malaysian study needs careful reading. At release the deviation climbs to about 60 degrees, a mean of 59.75 with arrows between 32 and 79, but the frame used to measure it is the tenth after the arrow leaves: a few tenths of a second later, when the drawing arm is already in its follow-through. It describes the moment after the shot more than the shooting position, and should not be confused with the anchor measurement.
Then there are the differences between archers. In 25 athletes from the Ukrainian national team, measured from above on video, the elbow angle ranged from 31.7 to 49.7 degrees, and the angle between forearm and shoulder line from 116.8 to 148.4: a spread of about eighteen and more than thirty degrees among athletes from the same national team. Many other measurements show the same, and the registry marks the idea that elite archers share a single correct technique as refuted: the skeleton of the shot is common, but angles, timings and muscles are largely personal.
Aligning the elbow and the score
This is where the picture splits. The only study that moved the alignment on purpose and then looked at the target is also the strongest in design: the 31 Korean archers were randomised, taking sex and experience into account, and 16 of them did eight weeks of lower trapezius exercises, while the other 15 served as controls. The alignment changed: in the exercise group the deviation fell from 10.06 to 7.94 degrees, while in the control group it went from 9.20 to 9.67. The score over 15 arrows at 70 m, on the other hand, stayed where it was, 132.13 before and 132.06 after.
The result allows two readings, and the registry entry keeps both. The first is that two degrees of alignment make no difference to the score. The second is that two degrees are too few to show anything over 15 arrows, in archers who already started from a national-squad geometry. What the study does allow us to say is narrower: in that group the alignment changed, and over those eight weeks the score did not follow it. The link between the lower trapezius and the line of force has its own entry too, currently contested, because a second programme for the same muscle, which measured the shoulder rather than the line, saw no change in arm position. Once the other studies of shoulder-blade exercises are counted, the effect of such exercises on the score is contested as well.
The Malaysian study is the only one to set the elbow deviation against the score arrow by arrow, 576 arrows in all, and it splits within itself. The authors separated the eight best archers, from 285 out of 360 upwards, from the eight weaker ones: at anchor the deviation was 4 degrees in both groups, and the measurement after release did not separate them either. Arrow by arrow, however, a slightly smaller deviation after release went with slightly higher scores, with a correlation of −0.083 and p = 0.047. A correlation this small clears the statistical threshold because there are so many arrows, not because the effect is large, and it accounts for less than 1% of the differences from one arrow to the next. At anchor, where the coaching rule wants the elbow in line, there was no link at all (−0.032, p = 0.442).
The rest of the literature points the other way or settles nothing. In eight high-level French archers at 70 m, abduction of the drawing shoulder and flexion of the elbow did not separate arrows scoring 10, 9 and 8 or less, either across the group or within any single archer. In seven university recurve archers at 10 m, the angle of the line of force was not significantly related to accuracy (r = −0.533, p = 0.218). The only systematic review that comes out in favour appears in a journal with doubtful editorial standards and leans on that very study of 576 arrows. With one small supporting result and several going the other way, the claim is contested.
Elbow and shoulder: experts and beginners
Then there is the question of height. The first comparison of elbow height in experts and beginners is a two-page Japanese letter from 1997 on thirty barebow archers filmed from the side with one camera: fifteen with at least three years of practice and fifteen with less than one. The authors measured the angle between the two shoulders and the drawing elbow, and read it as the height of the elbow: the wider the angle, the higher the elbow sits towards the shoulder line. It was about 155 degrees in the experts and about 140 in the beginners. The direction is the one archers are taught, and the entry stating that experts keep the string elbow more open records it as reported, with clear limits: the figures are means of three shots read off a flat image, with no statistical test at all, and nobody has linked that angle to the score.
The other side of the same question is the shoulder. In 30 Japanese recurve archers, split into three groups of ten by their score over 72 arrows at 70 m (633, 593 and 557.5 points on average) and filmed in the laboratory shooting at a target two metres away, the drawing arm sat at 99.0 degrees from the vertical in the elite as the clicker went off, at 105.8 in the middle group and at 107.2 in the lowest-scoring group; in the elite the elbow was also less flexed, 138.3 degrees against 143.3 and 143.6. Since 90 degrees means a horizontal arm, the better archers held the elbow closer to shoulder height, and the entry stating that higher-level archers keep the drawing arm less open at the shoulder is reported. None of the three groups reached the 90 degrees that the same study takes from the literature as the ideal position. In the 31 Korean elite archers the same angle sat at around 101 degrees, and the eight weeks of exercises did not move it: from 101.0 to 102.2 in the exercise group.
Put side by side, these figures do not describe an ideal height. The 1997 Japanese study says that in experts the elbow rises towards the shoulder line, the 2018 study that in the elite the arm is less open; a recent French thesis finds the elite with the drawing shoulder more abducted than intermediate archers. These are three different definitions, taken from different views and on different archers, and for now they cannot be added together. What remains is more modest: in the Japanese and Korean elite the drawing arm sat about ten degrees above the horizontal, and the 90-degree reference does not come from a measurement on strong archers.
Elbow height and tremor
On why height might matter there is a single laboratory study, the one the Japanese work on elite archers cites in support. Ten male beginners held six simulated positions for seven seconds with a 26-pound recurve: drawing arm at 90, 100 and 110 degrees of elevation, at partial and full draw, with motion sensors on the humerus and scapula. At 100 and 110 degrees tremor was greater than at 90, both in the 8 to 12 Hz band (p = 0.002) and in the 3 to 7 Hz band (p = 0.005), while shoulder muscle activity did not change with arm height (p = 0.53 for the humeral muscles, 0.29 for the scapular ones). In those ten, raising the elbow did not ask more of the muscles but made the position less steady.
In the same work the tremor was larger at the humerus than at the scapula, and only weakly linked to muscle activity, as the entry on the origin of drawing-arm tremor sums up. The limits lie in the sample and the task: beginners, positions held without shooting, no arrows and no score. That is why the registry counts those ten as the general population rather than as archers, and marks the claim that an elbow above shoulder height makes the hold less steady as transferred: on archers, so far, there is nothing. Whether the same happens to a trained archer during an end has not been measured.
The elbow and the back
Archers are taught to keep the elbow in line mainly because that geometry should pass the load to the back. Here the evidence is broader and points almost entirely one way, though it is thinner than the advice suggests, and the entry stating that expert archers hold the draw with the shoulder and back rather than the forearm remains reported. In the 30 Japanese archers, over the second before the clicker went off, the lower trapezius on the drawing side worked at 54.0% of maximal contraction in the elite, 39.4% in the middle group and 34.0% in the weakest, and the ratio between upper and lower trapezius fell from 2.33 to 1.23 as level rose. In a Turkish group of 23 people, elite archers, beginners and non-archers, the higher the score, the less the forearm muscles of the string hand worked around the clicker, with correlations reaching 0.74. That figure needs handling with care, because the non-archers were simply assigned a score and the time windows were chosen to maximise the correlation itself.
Two warnings keep the picture in proportion. The first is that the forearm does not get a rest: in five elite Russian archers the finger flexor of the string hand worked far harder than its counterpart on the other arm, with an asymmetry coefficient of 0.86 on a scale where zero means parity. The back carries the load, but it does not take over from the forearm. The second is that the link between elbow position and back work is less direct than the rule suggests: in seven Taiwanese university archers, compared with seven who did not follow the programme, twelve weeks of lower trapezius exercises halved the ratio between upper and lower trapezius on the drawing side, from 2.2 to 1.1, without significantly changing the shoulder angles. The balance between the parts of the trapezius can change without the elbow moving.
What I take from this for the range
Taken together, these studies do not ask us to drop the elbow correction, only to give it its proper weight. I take three things from them for the range, though none has been tested in an experiment on scores.
The first concerns the number of degrees. Chasing an elbow exactly on the line has no study behind it: the national-squad archers measured sit a few degrees away, between the 4 at anchor in the young Malaysians and the 10 in the Korean elite measured with another definition, and when an intervention removed two degrees the score did not change. A small deviation, within what is seen in strong archers, is not in itself a fault to be corrected.
The second concerns consistency, and it is a question more than an answer. No study has yet tested whether what matters is the alignment itself or how much it stays the same from one arrow to the next, and in the eight French elite archers the drawing-arm angles did not separate good arrows from bad ones even within the same archer. Anyone who wants to look into it can be filmed for a whole end, always from the same spot, ideally from behind and slightly above: comparing your own position from one arrow to the next tells you more than comparing it with an angle from a manual. I explain how to set up the recordings in the article on video analysis.
The third concerns height. For a beginner, holding the drawing arm around the horizontal or slightly above is a reasonable starting point, because it goes in the direction of the little data there is: the Japanese elite at 99 degrees and, in beginners in the laboratory, less tremor at 90 than at 100 or 110. For someone who has shot for years with a slightly higher elbow, on the other hand, no study says that lowering it improves shooting, and the Korean data suggest that the angle changes little even after weeks of targeted work.
What is missing
First of all, a common definition is missing: the same line of force, measured with the same landmarks, from the same view and in three dimensions, across more archers. On that basis we would need a study linking alignment to where each arrow lands, with statistics that take account of arrows from the same archer, and separating average alignment from its consistency. For elbow height, the tremor comparison would need repeating on trained archers, actually shooting and with the score alongside. Until then, the elbow-in-line correction rests on plausible geometry and a minimal association, and the only experiment that put it to the test did not see the score move.