{"id":"arxiv/2610.42058","url":"https://arxiv.org/abs/2610.42058","tags":["math.ST"],"score":0,"title":"Variance-adaptive Chernoff bounds: nonasymptotically valid, asymptotically sharp confidence sets for unbounded means","domain":"math.ST","source":"arxiv","authors":["Aaditya Ramdas","Diego Martinez-Taboada"],"reviews":0,"subject":"math.ST","summary":"Consider the classical statistical problem of sub-Gaussian mean estimation under martingale dependence: given $n$ observations with conditional mean $μ$ and a known conditional sub-Gaussian proxy $s>0$, we seek a nonasymptotically valid confidence interval for $μ$. The usual Chernoff interval has half-width $s\\sqrt{2\\log(2/α)/n}$; it neither adapts to the actual standard deviation $σ\\le s$ nor attains the Gaussian quantile $z_{1-α/2}$. We construct an analytically closed-form nonasymptotically-valid interval under martingale dependence whose endpoints, at every iid law with $σ> 0$, agree…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-30","reactions":{}}