Filming yourself while you shoot is the cheapest way to see your own shot from the outside, and the gap between what an archer feels they are doing and what the video shows is often wide. How much the filming itself raises your score is another matter, and research says less than people assume: the two archery studies that report a gain were published in journals with questionable editorial standards, and in both the footage was reviewed with a coach, so the registry treats the question as not studied. There are more precise data, if small ones, on what a camera reads well and what it reads badly: in three women archers, for instance, a camera placed diagonally behind them tracked the vertical sway of the aim and not the horizontal sway.
This guide covers how to film yourself and what to look for in the footage, where video measures badly and when a sensor gets further. The last part is about working on your own: how to compare the feel of a shot with a measurement, and what the only study on archers found about the training journal, one of the most repeated pieces of advice there is.
Phone, tripod and frame rate
You need very little: a smartphone that records at 60 fps or more, which covers almost every model of the last few years; a tripod or phone mount for 15 to 30 euros, because handheld footage moves more than the bow does; and an app that steps through frame by frame, such as OnForm, or simply the slow-motion mode of the phone's camera.
The frame rate depends on what you want to see. For posture and the pre-shot routine 60 fps is enough, because these movements last seconds. The release happens on another scale. In the finest measurement available, sixteen world-class recurve archers filmed at 4,000 frames per second, the string stays on the fingers for 1.9 milliseconds and the arrow leaves the string about fifteen milliseconds after the release begins; an older measurement, on an archer from the Italian national team, gives twice that, which is why the registry treats the exact figure as contested. The order of magnitude, a hundredth of a second, is the same for everyone, and it is enough to do the sums: at 30 fps there is one frame every 33 milliseconds, so the whole release falls between two images, while at 240 fps there is one every 4 milliseconds or so and the release takes up a few frames. Even then you will not see the fingers leaving the string, because that takes less than one frame, but you will see the hand clearly just before and just after, and that is what you need.
The four camera angles
-
Front
The camera sits at right angles to the line from the archer to the target, on the side the archer's chest is facing. It shows posture, the line of the shoulders and head position, and it sees the face from the draw-hand side, so it also gives a clear view of the anchor and the draw length, meaning where the point sits relative to the clicker. It is the basic angle for posture and anchor.
-
Rear
This camera is also at right angles to that line, but on the opposite side, behind the archer's back. It shows the shoulder blades and the symmetry of the shoulders, and it tells you whether the back or the arm is doing the work through the draw and expansion.
-
Side
The camera sits on the extension of the same line, beyond the archer, looking towards the target: it is called the side angle because the archer stands side-on to the target, so from here you see them from the side. It shows the line of force, meaning the alignment of arrow, draw hand and draw elbow, the line of the bow arm and the direction in which the hands move after the release. It is the reference angle for alignment.
-
Release close-up
The camera sits close to the draw hand and records at 240 fps. You see the hand and fingers just before and just after the release, and the path of the hand. It is for archers working on the details of the release.
For most archers the first three are enough; the fourth is for competitive archers working on the finer details of the release.
What to look for in the footage
-
Posture
From the front and rear angles
A level line across the shoulders, with the bow shoulder staying low through the draw; hips under the shoulders; an upright torso; the head turning towards the target without tilting.
Warning signs: the bow shoulder rising during the draw, the torso leaning towards the target or backwards, the head tilting or moving to meet the string.
-
Anchor
From the front angle
The hand arrives at the same point on every arrow, and the string rests on chin and nose the same way each time, or on the reference point of your own style if you use a different anchor.
Warning signs: an anchor that shifts by a few millimetres from one arrow to the next, the head moving to meet the string instead of the hand coming to the face, a string that touches the nose on some arrows and not on others.
-
Release
From the slow-motion close-up, then the front and side angles
The draw hand travels back along the neck, close to the face and in line with the arrow, and the string leaves cleanly. In high-speed footage of elite archers the fingers stay almost curled, but that is the effect, not an instruction to follow: in fifteen highly skilled archers, nearly all from the US national team, electromyography shows seven opening actively with the extensors, and their mean score for the year is almost identical to that of the other eight, 1,213 against 1,209 out of 1,440, and the registry treats the rule that the release must be purely passive as refuted. Slow motion also shows whether the string touches the face on its way out: among the sixteen world-class archers filmed at 4,000 frames per second it happened in four, with light contact, and the study does not say whether it costs points.
Warning signs: the hand flicking away from the face to the side, the hand moving forwards towards the target, the hand staying put at anchor instead of travelling back.
-
From the side and front angles
After the release the body moves far more than while aiming, above all the arms and the bow: the bow drops forwards into the sling and the draw hand travels back, and that is normal. What to watch is the direction of the movement, and how much the body comes apart: in thirty-nine elite archers on a force plate, a lower peak speed of the centre of pressure under the feet after the release went with higher-scoring arrows. Video cannot show that; what it does show, and what coaches usually ask for, is the bow arm staying up, towards the target, after the arrow has gone.
Warning signs: the bow arm dropping or swinging sideways as soon as the arrow leaves, the head lifting to follow it, the torso opening up.
What video measures badly
A single camera gives a flat image, and that is easy to forget once you start measuring. An angle read off the screen, such as the angle of the draw elbow, matches the real angle only if upper arm and forearm lie in a plane parallel to the camera; if the elbow moves towards the camera or away from it, the angle on screen changes even when the arm is still. That is why measurements can only be compared between recordings made from the same spot, at the same height and the same distance.
Aiming is the clearest case. In a comparison of three measuring systems, three women archers from the German national squad shot twelve arrows each at 30 metres, recorded by a single camera at 25 fps placed diagonally behind them, by a two-camera system and by an optoelectronic sensor mounted on the bow. Over the last second before the release the single camera tracked the vertical movement of the sight well, with correlations of 0.69 to 0.97 against the sensor, and did not track the horizontal movement at all, with coefficients between -0.31 and 0.25. The archer with the poorest horizontal agreement was also the one whose bow rotated most in the frontal plane: from behind, a rotation of the bow around the sight and a sideways shift look the same. That is three people, and the registry treats the claim as reported, but the practical consequence is simple: from footage taken behind the archer you can judge whether the sight rises and falls, not whether it drifts left and right.
Then there are the apps that draw a skeleton over the footage. That the model finds shoulders, elbows and wrists you can see by eye; how accurately it finds them during a shot, with the hand and string close to the face, has not been measured in a way that can be read. The only check of the error covers eight still photographs of a single archer, with a scale factor given only through examples, and the registry treats the question as not studied. While the error remains unknown, the automatically drawn lines are best used to see the shape of a movement, not to read off a degree.
Timing works better. On fifty-eight competition clips of world-class recurve archers, filmed from the front at 30 fps, a system based on pose estimation measures the time between anchor and release with a mean error of 0.34 seconds against an expert's manual measurement. It works because at anchor the angle of the drawing arm almost stops changing, and indeed relying on the string near the face alone pushes the error up to 1.4 seconds. Recognising every phase is harder: in another study, on twenty archers, a classifier tells three phases apart in 90% of cases in recurve and 76% in barebow, and gets confused precisely between drawing and aiming and the expansion, where posture changes very little. The registry treats it as reported that the phases of the shot can be recognised and timed automatically from video, but an error of a third of a second is fine for counting the timings of a match, not for judging an arrow.
Beyond video: sensors
Where video struggles, a sensor can do better, provided you know what it measures. A wristband with an accelerometer on the bow-hand wrist, tested on six adolescent archers in an official competition, separates the stretches of signal that contain a shot, from the start of the draw to the release, from those at rest. With the three raw axes it manages this in 84.5% of cases; adding two quantities calculated from the signal, total acceleration and its smoothed derivative, takes it to 96.0%. The sensor is the same in both cases, so the difference comes from the processing. It does not tell drawing, aiming and release apart one by one: it recognises the shot against everything else. It is a single study, released as a preprint, which does not show how the model copes with an archer it has never seen, and the registry treats it as reported.
The other use of sensors is a real-time signal. In a Japanese school club, an accelerometer and a compass were fitted to a beginner's bow; they record the tilt and direction of the bow on shots that land in the ten, and then signal, with a light next to the sight or with a sound, when the bow comes back within one degree of that position. With the signal, the average distance between the three arrows of an end fell from 17.3 to 10.7 cm at 30 metres with the light and from 11.9 to 5.3 cm at 18 metres with the sound, over twelve ends per condition. The paper does not say how many archers shot, probably just one, and there are no statistics: the registry treats the claim as reported, a promising case rather than proof that the group tightened because of the signal.
A monthly protocol
-
Standardised recording
One session a month set aside for filming: 12 arrows per angle, in normal training conditions, with the tripod always in the same spots. It is worth marking them on the ground, because only footage taken from the same spot can be compared.
-
Review at home
Without rushing, watch each arrow at normal speed and then in slow motion, compare the 12 arrows with each other to find what repeats, and finally set them beside the footage from previous months.
-
Work plan
Choose one or two technical elements to work on over the next four weeks, and the month after, reassess everything with new footage.
For a beginner one session a month is enough: you need time between recordings to work on what you saw, and closer sessions end up showing the same things. A competitive archer can film more often, even every session, provided the analysis afterwards is selective. The two most common mistakes are watching the footage in passing, for reassurance or out of dismay, and trying to correct everything you see at once, which tends to break the shot apart. That is why it pays to work on one element at a time and consolidate it before moving on to the next.
Keep the footage under names that give the date, angle and distance (for example 2026-06-15_side_70m.mov), in folders by year and month. Watching your shooting from six months or a year earlier shows the slow changes, the ones you do not notice from one day to the next.
Feel and measurement
Video is also useful for something archers rarely do: comparing what you felt with what happened. The idea has specific, if small, support. In a Ukrainian study from 1998, a device fitted to the clicker and the sight measured the time between the clicker dropping and the release; after every arrow the archer stated their own estimate of that time and compared it with the reading, over a programme of fifteen to twenty-two sessions. After seven to ten sessions the estimation error given is 5 to 8 thousandths of a second, and in the experimental group release timing becomes more regular and the competition result improves, while in the control group, according to the author, nothing changes. The number of archers is not stated, and those 5 to 8 thousandths sit in a table describing the structure of the method, so it is unclear whether they are a result or a target: the registry treats it as reported that comparing feel with measurement teaches archers to recognise their own release timing.
A 2016 case also shows that feel alone is not enough: an experienced archer with an inertial sensor on the bow judged some shots bad that the measurements showed to be good, and gave each judgement before knowing the score. Video lets you do something similar, and this is my own application, not what the study tested: straight after the arrow, say out loud how it felt (on time, late, snatched), then watch it back. With front footage at 240 fps that frames both the clicker and the hand, the time between the clicker and the string leaving is a few dozen frames, and you can count it without any instruments.
The training journal
Keeping a journal is one of the most repeated pieces of advice for archers who train on their own, and there is a single study on archers. Eight Singapore national team archers, recurve and compound, kept a guided self-reflection journal for five weeks: at the end six of the eight had a slightly lower mean over six 36-arrow rounds and two a slightly higher one, with changes between 6.8 points down and 1.5 up (from -2.31% to +0.45%), and no statistical test. Nobody filled it in completely, since the sessions written up range from 9 to 23 out of 30, and the archers rated it useful, 7.63 out of 10, mainly as a reminder, while complaining about the time it takes and about too many questions that were always the same.
The registry treats it as reported that a journal does not improve the score. With eight people, no control group and changes within the normal noise of a round of arrows, this is not proof that a journal does not help; it is, however, a reason not to make it compulsory. A couple of lines next to each session's footage, with the element you are working on and how the best and worst arrows felt, help you find things again and compare feel with the image.
Watching footage of champions
Comparing yourself with footage of high-level archers has some support, small, among young beginners. In twenty child archers aged 9 to 11, ten watched videos of shots by Italian national team athletes and ten did not: at the coaches' final assessment, drawing and aiming were rated good or excellent in 6 children out of 10 in the video group and in 2 out of 10 among the controls, starting from 1 in 10 in both. There is no statistical test, the groups were not randomised, the score on the target was not measured, and the release remained the hardest phase even for those who watched the videos; the registry treats the claim as reported.
For an archer who is already trained, the question is what to copy. Since 1990 every sample measured has found elite archers who are highly consistent with themselves and different from one another, in angles, muscles and timing, and in competition footage of the top fifteen men and top fifteen women in the world ranking every phase of the shot shows at least two styles. The registry treats the idea of one correct technique shared by champions as refuted, so copying a champion's outward form achieves little. What they share is a shot that repeats itself with great consistency arrow after arrow, and it is against that consistency that your twelve arrows are worth comparing, rather than against their form.
Reviewing footage with your coach
The biggest limit of video self-coaching is that you only see the problems you already know how to recognise: the unfamiliar ones stay invisible even when a trained eye spots them at once. That is why it pays to review the footage now and then with an experienced coach, in person or remotely, even just by sending it between lessons.
Among studies on archers, it is also the only use of video in which scores went up. In 31 Thai archers at national and international level, the score over 10 arrows at 70 metres went from 78.96 to 85.61 after a period of work with video and individual corrections from experts, with no control group; in 60 under-18 archers split in half, those who filmed themselves and reviewed the footage with their coach gained 4.23 points in a 40-metre test, against 0.27 for the group that trained as usual. Both studies, though, appeared in journals with questionable editorial standards, and a questionable venue never decides a status: the registry reports them, but treats as not studied both the idea that reviewing footage with your coach speeds up progress and the idea that filming yourself raises your score. Even taken as they stand, they say that an archer who watches their own shots with their coach improves, not that the video does it on its own.
A coach on the line remains irreplaceable for physically correcting posture and for stepping in at the moment; video adds the chance to replay the same instant as often as you need and to compare one month with another. Online you can gather opinions from many archers, but they often contradict each other, and a beginner struggles to tell qualified advice from casual opinion: one coach whose work you know is the better option.
Apps and software
OnForm lets you annotate, put two videos side by side and overlay grids and lines. For a basic analysis the phone's camera in slow motion and any player that steps frame by frame, such as VLC on a computer, are enough. Kinovea, free and open source, measures angles, paths and speeds, within the single-camera limits described above, and takes a little time to learn at first. It is the same kind of analysis I wanted to make more accessible when I built Eyercer, the software I use with my athletes.