This paper investigates the relative empirical performance of 3 stories of error' in decision-making experiments - finding that the constant-error-probability story does not fit particularly well, but that the white-noise and stochastic preference stories perform considerably better. The paper...