Posterior summaries with dependence

Topic: Markov chains, Gibbs sampling, and Metropolis–Hastings · more questions on this topic
‹ A long run and a missed mode Random-walk scale ›

Explain why the sample mean of retained MCMC draws can still estimate a posterior mean although the draws are dependent. Why is the nominal count L not the appropriate measure of information? Name three diagnostics or summaries that address this problem.

This exercise has no worked answer.