Many popular QC algorithms are probabilistic in nature, like Grover's, Shor's, QAOA ..etc
For some of these we have formulas that give probabilities of success (like for Grover's and Shor's), and for others that are harder to analyze we estimate these probabilities by testing the algorithm on various problem samples (like for QAOA).
Is there a general framework that we use to compare the performance of either type of probabilistic algorithms against deterministic algorithms to ensure a fair comparison?