The aim of this simulation study is to determine the effect of uncertainty in intracavitary probe electrode position on the accuracy of estimated endocardial potentials. Intracavitary probe position uncertainty is simulated by randomly moving an idealized probe surface about the center of an idealized left ventricular endocardial surface. These random deviations represent possible probe locations that are incorporated as correlated noise. An optimum inverse transfer coefficient matrix, relating intracavitary potentials to endocardial potentials, is computed and subsequently used to calculate the best linear estimate of the true endocardial potentials. For uncorrelated endocardial potentials and probe position uncertainty within 1.5 mm of the coordinates of the exact probe electrode locations, a root-mean-square (rms) error of 34.0% is obtained. Increasing probe position uncertainties to 3.0 and 6.0 mm results in rms errors of 60.8 and 88.3%, respectively. For endocardial potentials that are 90% dipolar, the rms errors for probe position uncertainties of 1.5, 3.0, and 6.0 mm are 11.3, 19.6, and 28.5%, respectively. These simulation results imply that position uncertainty of a multielectrode, intracavitary probe can be a major source of error in estimating endocardial potentials from intracavitary potentials.