Start detection in resisted sprinting:a validity and reliability analysis of encoder- and sensor-based methods /

This study evaluated the validity and reliability of start detection methods for determining 5-m sprint time using a resistance device with an encoder. Seventeen recreationally trained participants completed 12 resisted 5-m sprints across two sessions. Sprint start was identified using three encoder-based methods (position, velocity [MEv], and acceleration), a rear-foot release sensor and hip-marker kinematics (KIN) as a proxy for forward center-of-mass displacement. Reliability was evaluated using intraclass correlation coefficient (ICC), coefficient of variation (CV), typical error (TE), standard error of measurement (SEM), and minimal detectable change (MDC), while agreement was examined using Pearson correlations and Bland-Altman analysis. All methods showed moderate to excellent reliability (ICC = 0.54–0.93; CV ≤ 11.1%) and strong agreement with KIN (r = 0.90–0.94). MEv showed the smallest bias (−0.00 ± 0.07 s) and narrowest 95% limits of agreement, and is therefore preferred for sprint start detection in practical settings.