Dynamic accuracy of the state transition method in simulating linear systems