The segmental method for estimating the centre of mass (COM) location of the human body has been widely used since 1889. How closely this method agrees with direct measurements of the location and movement of COM during activity however, remains unclear. To test this, a novel reaction-board utilizing life sized projections of human subjects is designed for measuring COM location. Agreement between the segmental method and the more direct reaction-board measurement method is then assessed. Our data demonstrate that the reaction-board system has a physical maximum error of 1.28 cm and 1.95 cm for locating COM along the vertical (board length) and horizontal (board width) axes respectively, and show that the reaction-board and segmental methods agree to within limits of 6.0 cm for the location of COM and to within 5.6 cm for the movement of COM between two points, in recumbent individuals. Applied to running, the segmental method agrees to within limits of 4.8 cm for oscillation of COM and 5.3 cm for stride median COM height. The segmental method agrees with a more direct technique of known accuracy, the reaction-board method, most closely when measuring averaged oscillation over repeated strides, where it displays a measurement error range of 5.1 cm to 0.1 cm in runners.