Reasons for performing study: The relationship between the biomechanical horse–rider interaction and endurance race performance requires further investigation. Objectives: To characterise, both quantitatively and qualitatively, elite and advanced horse–rider dyads on the basis of the biomechanical horse–rider interaction during endurance races. Study design: Five elite and 5 advanced horse–rider dyads were recorded during CEI*/CEI** endurance races using 2 synchronised triaxial accelerometers each placed close to horse and rider centres of mass. Methods: For each horse–rider dyad, analyses focused on the vertical displacements of horse and rider per stride. This allowed quantification of the proportional use of each gait and riding technique per loop. The quality of the biomechanical horse–rider interaction was examined through the relative phases (RP) of their respective vertical displacement minima. Instantaneous speed and rider heart rates were recorded using a global positioning system device/heart rate monitor. Results: All dyads predominantly used 2 riding techniques per gait. The 2-point trot proportion was limited in both groups (11%). Throughout the race, the advanced horse–rider dyads showed a global stability in speed, in the proportion of 4 combinations of gait and riding techniques and in mean RP. However, the elite horse–rider dyads initially had higher mean RP values (P<0.01), and from mid-race to the end an increasing proportion of sitting canter, with associated increases in racing speed (P<0.001) and in mean heart rate (P<0.01). Intradyad RP variability in 2-point canter increased in both groups (P<0.01). Conclusions: Accelerometers are a valuable tool to follow the quantitative and qualitative trends of the biomechanical horse–rider interaction during international endurance races. The overall results emphasise the influence of the level of expertise on the adopted gait and riding techniques, thus influencing the racing speed. It remains to be established whether fatigue and/or strategy underlie our observations.