1: \begin{abstract} This short note investigates convergence of

2: adaptive MCMC algorithms, i.e.\ algorithms which modify the Markov

3: chain update probabilities on the fly.  We focus on the Containment condition

4: introduced in \cite{roberts2007coupling}.

5: We show that if the Containment condition

6: is \emph{not} satisfied, then the algorithm will perform very poorly.

7: Specifically, with positive probability,

8: the adaptive algorithm will be asymptotically less efficient then

9: \emph{any} nonadaptive ergodic MCMC algorithm. We call such algorithms

10: \texttt{AdapFail}, and conclude that they should not be used.

11: \end{abstract}