The James–Stein estimator is a biasedestimator of the mean of Gaussian random vectors. It can be shown that the James–Stein estimator dominates the "ordinary" least squares approach, i.e., it has lower mean squared error on average. It is the best-known example of Stein's phenomenon.