{"id":"arxiv/2610.17403","url":"https://arxiv.org/abs/2610.17403","tags":["stat.CO","math.ST","stat.ME"],"score":0,"title":"Overlap, Not Moments: Finite-Sample Theory for Bridge Sampling","domain":"stat.CO","source":"arxiv","authors":["Giorgio Micaletto"],"reviews":0,"subject":"stat.CO","summary":"Bridge sampling estimates of marginal likelihoods are routinely reported with a delta-method Monte Carlo standard error (MCSE). Existing theory is asymptotic and cannot say whether a given run is long enough to trust the estimate or its MCSE. We give finite-sample answers. The optimal bridge estimator fits a logistic regression that separates posterior from proposal draws, and its accuracy is governed by their overlap, within a factor of two of the classification's Bayes error. Because the estimating function is a sum of bounded terms, no moment conditions are needed, and all bounds hold…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-10-12","reactions":{}}