It’s called the Cross Entropy Method because when we are picking the leads, we are basically doing Maximum Likelihood Estimation (MLE). This is essentially the same as minimizing the cross entropy.