This paper outlines MCAPP (Monte Carlo Analysis of Probabilistic Programs), a technique to reason about numerical values for expected outputs in probabilistic programming languages with probabilistically approximately correct guarantees.