Author(s): GUIGNARD, B., KOMAR, J., KARCHER, C., Institution: UNIVERSITÉ LYON, UCBL-LYON 1, Country: FRANCE, Abstract-ID: 2168

INTRODUCTION:
During the last European Handball Championships, teams performed on average more than 50 shots per game [1]. Recently, embedding microsensors inside the ball has enabled ecological monitoring [1] of those shooting actions during competition; however, the information provided exclusively inform on the number of shots, where more contextual information could be derived to support performance analysis [2]. The aim of this study was to develop an automatic shot-detection algorithm based on ball kinematics, starting from peak ball acceleration [3] and additional features derived from positional tracking: temporal (time for the ball to enter the 6-m zone), spatial (direction and angle), and spatio-temporal variables (ball velocity).
METHODS:
Ball position (X-Y coordinates) was continuously recorded at 10 Hz using a local positioning system (Kinexon SafeTag, Munich, Germany [4]) during all matches of the 2022 Women’s EHF European Championship. Seven randomly selected games were manually annotated from video to obtain shot timecodes (ground truth). These timecodes were synchronized with tracking data to extract peak acceleration and associated variables. Thresholds (mean ± 1.96 SD) were calculated and applied to the full tournament dataset to automatically detect shots. Shooting location and velocity were retained as performance indicators.
RESULTS:
The algorithm detected 3,528 shots with a mean velocity of 21.82 ± 3.76 m/s and an average distance of 6.84 ± 1.47 m from the goal center. Most shots (n=1102) were executed centrally (3-m width between goal line and 9-m line). Shots were mainly performed between the range 17 and 28 m/s (84.5% of shots), with 430 slower shots (<17 m/s) and 117 faster shots (>28 m/s).
CONCLUSION:
Elite women’s handball shots can be reliably identified using a limited set of transparent temporal, spatial, and spatio-temporal features derived solely from ball data. Such information is essential for a detailed analysis of the sport’s constraints and for improving our understanding of performance determinants. Our findings also support the need for sport-specific female performance analyses, as male-based references are no longer appropriate [1]. Accordingly, we propose updated velocity categories for elite women’s handball: C1 (<19.5 m/s), C2 (19.5–22 m/s), C3 (22–24.5 m/s), C4 (>24.5 m/s).
References
1. Pueo B, Tortosa-Martínez J, Chirosa-Rios LJ, et al. Scand J Med Sci Sports 2022; 32(3): 588–597. 2. Tan JQJ, Robertson S, Komar J. Int J Sports Sci Coach 2025; 20(2), 858-868. 3. Adams M, David A, Hesse M, et al. In: Proceedings of the 6th International Workshop on Multimedia Content Analysis in Sports. 2023, pp.83–91. 4. Hoppe MW, Baumgart C, Polglaze T, et al. PLoS One 2018; 13(2): e0192708.