An examination is made of the errors generated when calculating the magnitude of the epicardial potential gradient using voltages recorded from three unipolar electrodes in a fixed triangular geometry on the epicardial surface. The error sensitivity of the gradient calculation increases significantly with decreasing electrode separation and increasing common-mode voltage. For example, a mean error in calculated gradient magnitude of 67% results when measuring a 5 V/cm gradient at 50 V common-mode with 2 mm interelectrode spacing and random 1% voltage error. Optimization of interelectrode spacing and common-mode voltage to reduce overall error is essential for accurate measurement of low magnitude epicardial gradients.