This study examined the interday and intraday reliabilities and validities of various sprint performance variables on a nonmotorized treadmill (NMT) over distances of 10, 20, and 30 m. After habituation, 12 male team-sport players performed 3 sprints on the NMT on 2 separate days and an assessment of overground running performance, separated by 24 hours. Measurements included sprint times, mean and peak sprint speeds, and step length and frequency. Data analysis revealed no significant mean differences (p > 0.05) between NMT variables recorded on the same day or between days. Ratio limits of agreement indicated that the best levels of agreement were in 20-m (1.02 ×/÷ 1.09) and 30-m (1.02 ×/÷ 1.07) sprint times, peak (1.00 ×/÷ 1.06) and mean (0.99 ×/÷ 1.07) running speed, and step length (0.99 ×/÷ 1.09) and frequency (1.01 ×/÷ 1.06). The poorest agreement was observed for time to peak running speed (1.10 ×/÷ 1.47). These reliability statements were reinforced by coefficients of variation being <5% for all the variables except time to peak running speed (11%). Significant differences (p < 0.05) were observed between NMT and overground sprint times across all distances, with times being lower (faster) by approximately 25-30% overground. The correlations between NMT and overground variables were generally modest (0.44-0.67), and optimal for time to cover 30 m on day 2. Our data support NMT ergometry as a reliable tool for most of the sprint performance variables measured and reveal that the fastest 30-m overground sprinters were likely to be identifiable via NMT ergometry.