Nondominated sorting and convex hull peeling are two algorithms for ranking multi-dimensional data sets. In recent years, J. Calder and collaborators have shown that the outputs of both algorithms can be approximated via PDEs when the sample size is large. In joint work with A. Bou-Rabee, we analyze the large-sample limit of a related scheme, called Pareto hull peeling, and show that it, too, is governed by a deterministic PDE. In the talk, I will explain the connections between all three algorithms and give the main ideas of the proofs.