Breaking Down the Beach Sprint: Reliability of Race-Segment Analysis with a Two-Camera System

Abstract

Purpose

Beach Sprint Rowing (BSR) is a young Olympic rowing discipline and there is limited scientific evidence on how the different race segments contribute to overall competition performance. To address this gap, video-based race analysis is warranted, however it depends on observational systems with high inter- and intrarater reliability. Therefore, the aim of this study was to investigate the interrater and intrarater reliability of a two-camera race-analysis system in BSR using real competition footage from the 2025 European and World finals.

Methods

A total of 455 race videos were analysed by three sports scientists across six predefined race segments: Sprint 1, Entry, Rowing 1, Turn, Rowing 2 and Sprint 2. Inter-rater reliability was assessed using ICC(2,1), while intra-rater reliability was evaluated using ICC(3,1). In addition, standard error of measurement (SEM), minimal detectable change (MDC), coefficients of variation (CV), maximal differences and Bland-Altman analyses were calculated.

Results

Excellent agreement was observed across all analysed race segments for both interrater (ICCs ranging 0.896 to 0.997) and intrarater (ICCs ranging 0.950 to 1.000) reliability analyses, indicating consistently high reproducibility of the applied race-segmentation approach. SEM and MDC values remained low across all segments (SEM: 0.14-0.59s; MDC: 0.38-1.65s), while CVs ranged 8.7% to 47.7%. Bland–Altman analyses demonstrated minimal systematic bias between raters. Slightly larger variability and wider limits of agreement (LoA) were observed for the rowing and turn segments, likely due to increased camera distance and difficulties in identifying exact transition points.

Discussion

The two-camera system showed excellent interrater and intrarater reliability under real competition conditions. Slightly higher variability in rowing and turn segments was likely caused by greater camera distance and less distinct transition points, but overall measurement error remained low.

Conclusion

The present findings demonstrate that the applied two-camera system provides a highly reliable and practically feasible approach for race-segment analysis under authentic competition conditions in the new Olympic rowing discipline.