Documentation

LeanPool.Feige.VlassisThomas.Main

Exact calibration for means of nonnegative random variables #

This is the public interface for the main theorem of Vlassis and Thomas, An Exact Distribution-Free Test for Means of Nonnegative Random Variables (arXiv:2607.08415). Its implementation includes the two-point reduction, chain calibration, exponential-transfer argument, boundary approximation, measurable two-point mixing, and normalization to coordinatewise means at most one.

The Vlassis--Thomas exact calibration theorem, uniformly over small-universe probability spaces.