Integrated autocorrelation time
= Integrated autocorrelation time
{title2=$\tau_{\rm int}=1+2\sum_{k\ge1}\rho(k)$}
For a stationary <Markov chain> and a scalar observable with summable <autocorrelation function> $\rho(k)$, the asymptotic variance of its sample mean is $\operatorname{Var}(h)\tau_{\rm int}/n$. Thus the <effective sample size of a Markov chain> is approximately $n/\tau_{\rm int}$. Different observables can have very different autocorrelation times.