Artículo de publicación ISIThis study aimed to compare the between-session reliability of three typically measured velocity variables (mean velocity [MV], mean propulsive velocity [MPV], and maximum velocity [Vmax]) to assess vertical jump performance. Totally, 23 men had their squat jump (SJ) and countermovement jump (CMJ) tested against five different loading conditions (17, 30, 45, 60 and 75 kg) during two consecutive weeks. The two sessions of each jump type were performed within the same week separated by 48–72 h. The main finding was a significant difference in reliability between the variables, which were ranked from the highest to the lowest reliable as follows (median coefficient of variation [CV] and range): Vmax (CV = 2.35% [1.85%–3.23%]) >MV (CV = 3.29% [2.18%–4.40%]) >MPV (CV = 3.69% [2.08%–5.17%]). A significant variable × exercise interaction was also observed showing that the differences in reliability between the variables were meaningful during the SJ (MV: CV = 3.93% [3.06%–4.40%], MPV: CV = 4.61% [4.07%–5.17%], and Vmax: CV = 2.14% [1.85%–2.71%]), while no significant differences were observed for the CMJ (MV: CV = 2.43% [2.18%–3.70%], MPV: CV = 2.71% [2.08%–3.63%], and Vmax: CV = 2.40% [1.97%–3.23%]). These results suggest that the Vmax should be the recommended variable for obtaining a reproducible measure of lower-body ballistic performance, especially during the SJ exercise.