Introduction & Purpose The idea of some athletes having high strength levels but being too slow for a particular movement while others are fast but not strong enough for the demands of their sport exists among many coaches. Therefore, planning decisions are often based on the measurement results of speed and force capabilities and the interest in the relationships between additional load, force, velocity and power during the required movement is high. The luge start is a ballistic movement that requires both high forces and velocities to reach maximal performance (i.e. fast starting times). Examining the load, force, velocity and power capabilities of elite luge athletes during the luge start could deliver insights for training prescription to improve individual weak points and improve performance. This study investigated the relationship between the calculated maxima of force-velocity (FV), load-velocity (LV) and power-velocity (PV) profiles during a simulated luge start and on-ice start times in elite luge athletes. Methods Eleven Austrian national luge team athletes (1 female, 10 male, 25 [4.4] y, 182.4 [7.3] cm, 86.65 [10.4] kg) participated in this study. Measurements were conducted during a single testing session on a specially designed start simulator. Each athlete performed starts against 5 increasing resistances relative to their bodyweight (0%, 25%, 50%, 75%, 100%). For each pull, relative maximum force (Fmax), maximum power (Pmax) and maximum velocity (Vmax) were calculated using the measurements of a linear position transducer (GymAware PowerTool). An experimental approach was used to create the FV-, LV- and PV-profiles showing the highest R2. On-ice start times [s] were measured by photoelectric sensors (ALGE). A correlation matrix (p < 0.05) was used to analyze the relationships between the theoretical Pmax, Fmax (i.e. theoretical isometric maximum force), theoretically unloaded Vmax (Vmax(F0)), theoretical relative maximum loads (Lmax(V0)) and starting times obtained from the profiles. Results Ten linear FV profiles and 11 exponential LV were created (Figure 1; FV: R2mean = 0.9532 ± 0.06; LV: R2mean = 0.9918 ± 0.01). There were significant negative relationships between Vmax(F0) and on-ice starting times (r(9) = -.74, p < .05; r(9) = -.71, p < .05; r(9) = -.74, p < .05) and significant positive correlations between Lmax(V0) and ice start times (r(9) = .71, p < .05; r(9) = .71, p < .05; r(9) = .72, p < .05). Discussion The FV relationship during a simulated luge start can be represented mostly through a linear regression line, supporting the findings of former studies which showed a quasilinear relationship between force and velocity in multi-joint movements (Bobbert, 2012; Jaric, 2015). In line with earlier approaches, PV relationships could be represented quite accurately by use of second-degree polynomial regressions (Allison et al., 2013; Padulo et al., 2017). Contradictory to the linear LV relationships found in other movements (García-Ramos et al., 2018), the luge start LV profile can be better represented by exponential regression curves which could be due to the minimized friction in the luge start setup. Conclusion The abilities of accelerating high external loads and producing high velocities both affect luge start time.