{"id":"arxiv/2610.35973","url":"https://arxiv.org/abs/2610.35973","tags":["cond-mat.stat-mech","math.PR"],"score":0,"title":"Bayesian Inference in Partially Observed Critical Systems","domain":"cond-mat.stat-mech","source":"arxiv","authors":["Mihika Singhania"],"reviews":0,"subject":"cond-mat.stat-mech","summary":"We ask how much of a large random system must be observed before a global average can be inferred with controlled uncertainty. For the random transposition shuffle, a conjugate Bayesian estimator of the average displacement, computed from $m$ uniformly sampled cards, has relative uncertainty $1/\\sqrt{2m}$, independent of the deck size $n$. The law of total variance exposes a second error, the floor $R_n$, which arises because the observed configuration is a single draw from equilibrium: no relative tolerance below $κ\\sqrt{R_n}$ can be certified, even by a census. For the shuffle we derive…","category":"research","comments":0,"disputed":false,"division":"physics","mentions":0,"discussed":0,"published":"2026-10-26","reactions":{}}