1: \begin{abstract}

2: Let $(X_t)$ be a discrete time Markov chain on a general state space. It is well-known \cite{NT82,MT93,RR04} that if $(X_t)$ is aperiodic and satisfies a \emph{drift and minorization} condition, then it converges to its stationary distribution $\pi$ at an exponential rate. We consider the problem of computing upper bounds for the distance from stationarity in terms of the drift and minorization data.

3: 

4: Baxendale \cite{B05} showed that these bounds improve significantly if one assumes that $(X_t)$ is reversible with nonnegative eigenvalues (i.e.\ its transition kernel is a self-adjoint operator on $L^2(\pi)$ with spectrum contained in $[0,1]$). We identify this phenomenon as a special case of a general principle: for a reversible chain with nonnegative eigenvalues, any \emph{strong random time} gives direct control over the convergence rate. We formulate this principle precisely and deduce from it a stronger version of Baxendale's result.

5: 

6: Our approach is fully quantitative and allows us to convert drift and minorization data into explicit convergence bounds. We show that these bounds are tighter than those of \cite{R95,B05} when applied to a well-studied example.

7: \end{abstract}

8: