{"id":"arxiv/2603.09629","url":"https://arxiv.org/abs/2603.09629","tags":["math.ST"],"score":0,"title":"On the last time and the number of times an estimator is more than epsilon from its target value","domain":"math.ST","source":"arxiv","authors":["Nils Lid Hjort","Grete Fenstad"],"reviews":0,"subject":"math.ST","summary":"Suppose $\\widehatθ_n$ is a strongly consistent estimator for $θ_0$ in some i.i.d. situation. Let $N_\\varepsilon$ and $Q_\\varepsilon$ be respectively the last $n$ and the total number of $n$ for which $\\widehatθ_n$ is at least $\\varepsilon$ away from $θ_0$. The limit distributions for ${\\varepsilon}^2 N_\\varepsilon$ and ${\\varepsilon}^2 Q_\\varepsilon$ as $\\varepsilon$ goes to zero are obtained under natural and weak conditions. The theory covers both parametric and nonparametric cases, multi-dimensional parameters, and general distance functions. Our results are of probabilistic interest,…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-03-10","reactions":{}}