1: \begin{abstract}

2: The recently introduced backward Monte-Carlo method~[Johan Carlsson,\linebreak

3: \texttt{arXiv:math.NA/0010118}] is validated,

4: benchmarked, and compared to the conventional, forward Monte-Carlo method by

5: analyzing the error in the Monte-Carlo solutions to a simple model equation.

6: In particular, it is shown how the backward method reduces the statistical

7: error in the common case where the solution is of interest in only a small

8: part of phase space. The forward method requires binning of particles, and

9: linear interpolation between the bins introduces an additional error. Reducing

10: this error by decreasing the bin size increases the statistical error. The

11: backward method is not afflicted by this conflict. Finally, it is shown how

12: the poor time convergence can be improved for the backward method by a minor

13: modification of the Monte-Carlo equation of motion that governs the

14: stochastic particle trajectories. This scheme does not work for the

15: conventional, forward method. \\[4mm]

16: 

17: \noindent\emph{PACS:} 02.70.Lq; 02.60.Lj \\[-1mm]

18: 

19: \end{abstract}

20: