Let be a random variable and another independent and identically distributed random variable. We redefine the variance of to be the expectation of the energy between and :
This agrees with the usual definition of variance since
Given a sample , each pair has a mutual energy . We can then estimate the variance to be the mean of the pairwise energies which we also take to be the redefinition of the sample variance,
This agrees with the usual definition of sample variance since
and gives an explanation why we divide by in the final expression.