Objectives: This study aimed to determine the reliability of 10 and 20 km cycling time trial (TT) performance on the Velotron Pro in recreational cyclists, runners and intermittent-sprint based team sport athletes, with and without a familiarisation. Design: Thirty-one male, recreationally active athletes completed four 10 or 20 km cycling TTs on different days. Methods: During cycling, power output, speed and cadence were recorded at 23 Hz, and heart rate and rating of perceived exertion (RPE) were recorded every km. Multiple statistical methods were used to ensure a comprehensive assessment of reliability. Intraclass correlations, standard error of the measurement, minimum difference required for a worthwhile change and coefficient of variation were determined for completion time and mean trial variables (power output, speed, cadence, heart rate, RPE, session RPE). Results: A meaningful change in performance for cyclists, runners, team sport athletes would be represented by 7.5, 3.6 and 12.9% improvement for 10 km and a 4.9, 4.0 and 5.6% for 20 km completion time. After a familiarisation, a 4.0, 3.7 and 6.4% improvement for 10 km and a 4.1, 3.0 and 4.4% would be required for 20 km. Conclusions: Data from this study suggest not all athletic subgroups require a familiarisation to produce substantially reliable 10 and 20 km cycling performance. However, a familiarisation considerably improves the reliability of pacing strategy adopted by recreational runners and team sport athletes across these distances.