Optimising body position on the bicycle may improve performance and reduce overuse injury risk. However, effectiveness of bicycle configuration in preventing overuse injuries is unknown as methods to assess body position on the bicycle (usually via static assessment) lack accuracy and are limited to road cyclists. The aims of this thesis were to: 1) assess validity and reliability of methods for determining body position (kinematics) and pedalling kinetics in cycling; and 2) to analyse effects from changes in saddle height and horizontal position on pedalling kinetics and kinematics of cyclists and triathletes. Pedal force effectiveness, crank torque, joint angles and joint mechanical work during cycling on an ergometer were measured for recreational and competitive road cyclists and competitive triathletes, with and without lower limb pain, in five experimental studies. Analysis of static 6 o’clock crank position was not valid therefore body position on the bicycle should be measured using video from cycling motion. Instrumented pedals should be used for pedalling kinetics assessment as bilateral asymmetry was underestimated using the SRM® torque analysis system. Reliability of pedal forces from instrumented pedals computed seven days apart was high. Technique training using feedback from instrumented pedal forces at training workloads and high pedalling cadence improved power output. Using improved knee joint modelling, cyclists with and without pain had similar adaptations to changes in saddle height. Saddle height changes of 5% of the self-selected preferred saddle height did not result in substantial changes in pedalling kinetics. Moving forward or backward on the saddle affected joint mechanical work but not knee joint forces. Saddle position changed pedalling technique similarly in cyclists and triathletes. To optimise cyclists and triathletes body position on the bicycle and their technique, instrumented pedals should be used to assess pedalling kinetics and video analysis should be used to assess kinematics.