We use the numerical optimization techniques of Uskov et al. [PRA 81, 012303 (2010)] to investigate the behavior of the success rates for KLM style [Nature 409, 46 (2001)] two- and three-qubit entangling gates. The methods are first demonstrated at perfect fidelity, and then extended to imperfect gates. We find that as the perfect fidelity condition is relaxed, the maximum attainable success rates increase in a predictable fashion depending on the size of the system, and we compare that rate of increase for several gates.