This study was used to investigate the effects of familiarisation on a countermovement jump (CMJ) performed with hand held dumbbell accentuated eccentric loading (AEL) at 20% of body mass (CMJAEL20). Twenty‐seven adolescent males performedCMJAEL20 on three separate occasions. Statistical parametric mapping (SPM) detected significant differences in normalised force‐time data between session one and two (50%–95% of movement time), two and three (47%–48%) and one and three (66%–96%), but not in velocity‐ or displacement‐time data. Propulsion mean vertical ground reaction force (vGRF) had excellent reliability (CV% upper CI95 = 5.12–9.33; ICC lower CI95 = 0.99), whereas jump height exhibited good relative reliability (ICC lower CI95 ≥ 0.94) but moderate to poor absolute reliability (CV% upper CI95 = 6.72–15.36). Unweighting time and braking time showed moderate to poor reliability (CV% upper CI95 = 14.22–37.06; ICC lower CI95 = 0.46–0.89). Mean bias between sessions was ≤ 10% for all variables according to repeated measures Bland‐Altman analysis; however, fixed bias was observed in braking mean vGRF and propulsion mean velocity. Jump height, braking mean vGRF, propulsion mean vGRF and propulsion mean velocity exhibited good to acceptable limits of agreement (LOA; ≤ 20%), whereas all other variables were classified as ‘poor’ (>20%). Proportional bias was identified in unweighting vGRF%, braking mean vGRF and braking mean velocity. These findings suggest that although more than three familiarisation sessions may be required for unweighting and braking CMJAEL20 variables, reliable propulsion data, including jump height, were observed from session one.