 Asymptotic mean squared error

The asymptotic mean squared error is the leading large-sample approximation to the sum of variance and squared bias of an estimator. Minimizing it balances stochastic variation against systematic error; the resulting choice must still satisfy the assumptions used in the approximation.