Equine back pain is prevalent among ridden horses and is often attributed to poor saddle fit. An alternative explanation is that saddle fits are technically good but fit to the wrong configuration. Saddles are fit for the standing horse, but much of the time ridden is instead spent locomoting when the back experiences the greatest peak forces. We used an array of cameras to reconstruct the surface of the back and its movement during trot, walk and standing for five horses. We verified the setup’s accuracy by reconstructing a laser-scanned life-sized model horse. Our reconstructions demonstrate that saddles sit within a large, relatively low-mobile region of the back. However, saddles do sit adjacent to the highly mobile withers, which demands care in positioning and design around this important region. Critically, we identified that saddle curvature between standing and moving horses is substantially different, where trotting and walking horses have flatter backs than their standing configurations. Saddles designed around the locomoting configuration of horses may improve horse welfare by being better fit and decreasing the focal pressures applied by saddles.