OBJECTIVES: To examine the effects of cycling at preferred, forward and backward saddle positions on patellofemoral compressive and tibiofemoral compressive and shear forces. DESIGN: Cross-sectional. SETTING: An incremental cycling test to exhaustion determined cyclists' maximal aerobic workload and second ventilatory threshold. In a second session, 1-min cycling trial at maximal aerobic workload then three 2-min trials at second ventilatory threshold workload at preferred, forward and backward saddle positions. Right pedal force via instrumented pedals, lower limb joint kinematics via video and inverse dynamics were used to calculate knee joint forces. PARTICIPANTS: Twenty-one competitive cyclists (28 ± 7 years). MAIN OUTCOME MEASURES: Patellofemoral compressive, tibiofemoral compressive and shear forces, and knee flexion angle. RESULTS: Changes to forward/backward saddle positions did not substantially affect compressive forces for patellofemoral (1-4%) or tibiofemoral (1-3%) joints. Tibiofemoral shear force increased in backward compared with preferred (19%) or forward (26%) saddle positions. Knee flexion angle at 3 o'clock (22%) and 6 o'clock crank positions (36%) increased at the forward compared to the backward saddle position. CONCLUSIONS: Small increases in knee flexion angle (5-6°) explained trivial differences in patellofemoral and tibiofemoral compressive forces. Tibiofemoral shear force may be more sensitive to changes in knee joint angle compared to other knee force components.