A regression method for the power-duration relationship when both variables are subject to error.

GM provided equivalent results with respect to WLS in fitting the critical power model to experimental data and for its conceptual characteristics must be preferred wherever concerns on the precision of P measurement are present, such as for in-field power meters.