Matteo Campagna
ITEN
Menu
Menu

A wrist accelerometer can pick out the moment of the shot, and does it better when the signal is processed

1 study · 6 archers measured

ForContextAgainst202420262025, for, n = 62025, context, n = 6
Each circle is a study: its position gives the year, its row whether it is for, against or context, its size how many participants it measured. Light circles are studies on other populations or without archers.

A single 2025 study says so, released as a preprint and not yet through a journal. Six adolescent professional archers, four girls and two boys aged 14 to 16, shot in an official competition wearing a wristband developed by the authors on the bow-hand wrist, recording acceleration at 20 hertz, and a neural network had to separate the stretches of signal that contain a shot, from the start of the draw to the release, from those at rest: it does not tell the draw, the aim and the release apart, it picks out the shot from everything else. Given only the three raw axes, its accuracy is 84.5%; 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 eleven and a half points of difference come from the processing, and that is the part of the claim the data supports best. The limit lies in the numbers: six archers, a 70:30 split of the data that does not show how the model behaves with an archer it has never seen, and a second measure, which the authors designed for real use, that only rises from 67.9% to 75%.

Also phrased as: "smartwatch for archery"; "wrist sensor for archery"; "accelerometer shot detection"; "wristband that recognises the shot"

In practice

You hear that a smartwatch or a wristband is enough to count the shots and know how long each one lasts, as if the sensor could read the shot on its own.

Evidence

  • 2025
    For Archers Published later

    6 Chinese adolescent professional archers, 4 girls and 2 boys aged 14.9 to 16.3, in official competition, with a 20 Hz accelerometer wristband on the bow-hand wrist; 4-second windows classed as shot, from draw to release, or rest. LSTM network with the X, Y and Z axes alone: 84.53% accuracy, F1 83.45. Adding total acceleration and the smoothed derivative: 95.99% and F1 95.95. The gain comes from processing the signal.

    A Machine Learning-Based Multimodal Framework for Wearable Sensor-Based Archery Action Recognition and Stress Estimation · Xianghe Liu, Jiajia Liu, Chuxian Xu, Minghan Wang, Hongbo Peng, Tao Sun, Jiaqi Xu

  • 2025
    Context Archers Published later

    In the same study, the limits: sample-level accuracy, the measure the authors put forward for real use, only rises from 67.86% to 75.00%; a bidirectional network reaches 96.80% but costs more computing; the data are split 70:30 between training and testing, with no check on archers left out of training. The wristband was developed by the authors themselves, and the paper is a preprint.

    A Machine Learning-Based Multimodal Framework for Wearable Sensor-Based Archery Action Recognition and Stress Estimation · Xianghe Liu, Jiajia Liu, Chuxian Xu, Minghan Wang, Hongbo Peng, Tao Sun, Jiaqi Xu

Status history

  1. 2025 Reported The first and only study arrives: six adolescent professional archers in official competition with a 20-hertz wristband on the bow-hand wrist, and a neural network that has to separate the stretches of signal containing a shot from those at rest. With the raw axes alone it manages 84.5%, and adding total acceleration and the smoothed derivative takes it to 96.0%: eleven and a half points that come from the processing, since the sensor is the same. Six subjects, a preprint and a methodological validation with no test on new archers: reported.

Related

Related articles

Research lines

Cite
APA

Campagna, M. (2026). A wrist accelerometer can pick out the moment of the shot, and does it better when the signal is processed (Version 1). In Evidence Registry. Retrieved 16 October 2026, from https://matteocampagna.com/en/registry/equipment/wrist-accelerometer-detects-the-shot

BibTeX
@misc{campagna2026_accelerometropolsofasitiro,
  author = {Campagna, Matteo},
  title = {A wrist accelerometer can pick out the moment of the shot, and does it better when the signal is processed},
  year = {2026},
  howpublished = {Evidence Registry, matteocampagna.com},
  url = {https://matteocampagna.com/en/registry/equipment/wrist-accelerometer-detects-the-shot},
  note = {Version 1; accessed 2026-10-16}
}

Other claims in this domain

See all claims on equipment and tuning