Effective sample size of a Markov chain
ID: effective-sample-size-of-a-markov-chain
For a stationary Markov chain observable with positive variance and absolutely summable autocorrelation function ,If the limiting integrated autocorrelation time , the leading variance is . Matching this to the sample-mean variance of independent observations defines .
This is an observable-specific variance comparison, not a literal count of independent retained states. Positive serial dependence reduces effective sample size; negative dependence can make it exceed . For a stationary AR(1) observable with coefficient , and the asymptotic effective sample size is . Estimated effective sample size also requires reliable estimation of the long-run covariance; it does not prove that all modes have been explored.
New to topics? Read the docs here!