where we use an result borrowed from probability theory for the approximation: consider random variables