Yes indeed, I call it the revenge of statistics! Pseudoreplication rears its ugly head once again. Nonparametrics are certainly useful, but once you know things about the DGP, it’s often possible to do something “traditional” that’s better. Probabilistic programming, for life!