Let x1,…,xn be given real numbers with 0<m≤xi≤M for each i∈{1,…,n}. Let X be the discrete random variable uniformly distributed on {x1,…,xn}. The mean μ and the variance σ2 of X are defined as
μ(X)=nx1+…+xn and σ2(X)=n(x1−μ(X))2+…+(xn−μ(X))2.
By X2 denote the discrete random variable uniformly distributed on {x12,…,xn2}. Prove that
σ2(X)≥(2M2m)2σ2(X2).