1: \begin{abstract}

2:   The nonequilibrium work fluctuation theorem provides the way for calculations

3:   of (equilibrium) free energy based on work measurements of nonequilibrium,

4:   finite-time processes and their reversed counterparts by applying Bennett's

5:   acceptance ratio method. A nice property of this method is that each free

6:   energy estimate readily yields an estimate of the asymptotic mean square

7:   error. Assuming convergence, it is easy to specify the uncertainty of the

8:   results. However, sample sizes have often to be balanced with respect to

9:   experimental or computational limitations and the question arises whether

10:   available samples of work values are sufficiently large in order to ensure

11:   convergence. Here, we propose a convergence measure for the two-sided free

12:   energy estimator and characterize some of its properties, explain how it

13:   works, and test its statistical behavior. In total, we derive a convergence

14:   criterion for Bennett's acceptance ratio method.

15: \end{abstract}

16: