English

Rao-Blackwellized e-variables

Statistics Theory 2025-12-19 v1 Probability Methodology Statistics Theory

Abstract

We show that for any concave utility, the expected utility of an e-variable can only increase after conditioning on a sufficient statistic. The simplest form of the result has an extremely straightforward proof, which follows from a single application of Jensen's inequality. Similar statements hold for compound e-variables, asymptotic e-variables, and e-processes. These results echo the Rao-Blackwell theorem, which states that the expected squared error of an estimator can only decrease after conditioning on a sufficient statistic. We provide several applications of this insight, including a simplified derivation of the log-optimal e-variable for linear regression with known variance.

Cite

@article{arxiv.2512.16759,
  title  = {Rao-Blackwellized e-variables},
  author = {Dante de Roos and Ben Chugg and Peter Grünwald and Aaditya Ramdas},
  journal= {arXiv preprint arXiv:2512.16759},
  year   = {2025}
}

Comments

19 pages

R2 v1 2026-07-01T08:31:53.686Z