There are two competing estimators for ̂σ²:
̂σ²_MLE = 1/n Σ(Xi - X̄)²
vs
S² = 1/(n-1) Σ(Xi - X̄)²
Find their expected values. Are they unbiased?
Find their variances.
Find the relative efficiency of the two estimators, i.e., eff(̂σ², S²). Which estimator is better in terms of MSE? What if n → ∞?