How arrow speed is measured

Three numbers with one name

In a study, "arrow speed" can mean three things: launch speed, taken a few metres from the bow; the average over the whole flight, that is the distance divided by the time of flight; and the speed at one point along the trajectory. They do not match, because the arrow slows down on the way, and over a 70 m field it loses around 15% of its speed (the registry entry on speed lost in flight is "corroborated"). That is why the figure of just over 50 metres per second measured at 70 m on elite archers, an average over the full flight, sits below the roughly 60 at which a recurve arrow leaves the bow: both figures are in the entries on arrow speed at 70 metres and on the competition arrow.

The optical chronograph

The most common instrument is the chronograph: two light gates a known distance apart, with the speed given by that distance divided by the time between the two passes. The oldest light-gate measurement in the registry dates from 1985: at the Royal Ordnance range at Enfield, two gates three and six metres in front of the archer compared replicas of ancient bows with modern ones, recording the highest speed for each projectile rather than the average. Since then it has nearly always stood close to the bow: one metre from the rest on the 2009 shooting machine, two metres away in Meyer's work on the compound, and at one and six metres with the bows tested in Liège.

There is a single test of its accuracy. In 2011 a group validating an electronic arrow point shot a compound from a machine at the ballistics range of NIST, the American metrology institute, with the reference chronograph and three commercial chronographs on the same line. The reference gave a mean of 82.575 metres per second and the commercial units 81, 81 and 81.33: between 1.5% and 1.9% lower, all on the low side and all to the nearest whole metre per second, a resolution that at that speed is worth more than 1% on its own. The authors add that these instruments usually claim 0.5% and depend heavily on the lighting, and that beyond 20 m it is hard to get the arrow over the sensors. With four shots, one bow and three unnamed instruments, the entry on optical chronographs stays at "reported".

Park moved a chronograph out to 90 m, but only with a machine and in still air, because the arrow had to pass through a window of about 50 mm, and he gives an accuracy of about 0.3 metres per second. The offsets measured at NIST are the same size as the differences many comparisons are looking for: between the bare shaft and the slowest vane on the 2009 machine there were 3.2 feet per second out of 215, or 1.5%, and in Liège many fletchings were separated by a few percentage points of energy lost over five metres. An offset that always runs in the same direction matters little when two set-ups go through the same instrument under the same light; the resolution of the reading does not go away.

Sound and time of flight

A microphone behind the archer picks up the string and then the impact on the target. Heller used it to work out the average speed over 18 metres, correcting for the sound travelling back, and a 2016 study applied the same set-up to fifteen archers. The method tracks the variation from shot to shot well, the absolute values less so: in Heller's study the speeds are printed in metres per second with values around 200, which only make sense in kilometres per hour. Meyer placed two fixed microphones for 25 shots between 27.4 and 82.3 metres, beyond the chronograph's reach. A 2007 method uses the Doppler effect on a whistling point, with one shot per arrow and no chronograph alongside.

The Turkish field chronometer times the flight between a 150 g sensor on the grip and a vibration detector on the target: on twenty elite archers at 70 m it gives 1.290 seconds and 184.19 kilometres per hour, and the value holds a week later. Nobody has checked it against a chronograph, though, and 70 metres divided by 1.29 seconds would give somewhat more. Spratford and colleagues read the time of flight from footage of the target at 200 frames per second, one frame every five milliseconds: 1.24 seconds.

Cameras, radar and sensors inside the arrow

In the Japanese laboratory two cameras at 2,800 frames per second, 13 metres from the shooting position and 45 metres further on, measured speed and slowdown, correcting for the scale of the image, which changes with the distance from the camera. A program written in 2022 reads footage at 3,000 frames per second on its own and stays within 0.9% of manual reading horizontally, but the benchmark is another reading of the same images, not an independent instrument (more on this on the page on high-speed video).

A radar behind the archer works outdoors too, but in the 2012 Malaysian study some readings were lost to interference from teammates shooting alongside. The 2011 electronic point came within 0.59% of the NIST reference, with a standard deviation of 0.067 metres per second against 0.448 for the reference itself, over four shots. An inertial sensor in the shaft measured about 8% of speed lost over 55 metres; in kyudo a magnet in the shaft passing between two coils gives the speed just after the release.

When the instrument is not stated

In these cases the first check is the unit. The speeds of the 61 archers at the 2025 World Cup stage in Shanghai come from the event's data and are printed in metres per second: 292.0 for compound and 216.7 for recurve, values that only make sense in feet per second, about 89 and 66 metres per second. The entries on draw weight and on compound arrows outpacing recurve ones rest partly on this study. The Indonesian study on compound archers describes no instrument at all, and one from 2026 names paper tuning as its instrument for arrow speed, with no unit, whereas a sheet of paper shows how the arrow leaves the bow, not how fast it travels.

The instrument and the variation it has to detect

In archers the speed changes very little: in Heller's seven German archers the coefficient of variation lay between 0.23% and 0.58%, the Japanese archer repeated her launch speed within 0.6% against 0.3% for the compressed-air launcher, and the archer filmed in 2022 had standard deviations of 0.34 and 0.40 metres per second. These variations are smaller than both the offset and the resolution of the chronographs tested at NIST, and the measurements behind the entry on an archer's arrow speed consistency come mainly from a microphone, cameras and the NIST reference chronograph. On the target, though, they show: with a compound launching at 74.5 metres per second, one metre per second more at launch raises the impact by about 13 centimetres at 70 metres, according to the calculation in the entry on vertical spread. A chronograph that reads to the nearest whole metre per second cannot separate variations this small.

What to check

• which speed: at launch, averaged over the flight or at one point, and how far from the bow;
• the instrument, and whether it was checked against a reference;
• the resolution of the reading compared with the difference being sought;
• the unit, and whether the values are plausible;
• who is shooting, a machine, an archer or a launcher, and how many shots;
• the light, the wind and where the arrow passes over the sensors.

How design and sample are weighed is covered in How we read a study, the machines that shoot in place of the archer on the page on shooting machines, and what a few extra pounds change in the article on draw weight and your score.

@misc{campagna2026_enresearchmethodsmeasuringarrowspeed,
  author = {Campagna, Matteo},
  title = {How arrow speed is measured},
  year = {2026},
  howpublished = {Research, matteocampagna.com},
  url = {https://matteocampagna.com/en/research/methods/measuring-arrow-speed},
  note = {accessed 2026-10-16}
}