Showing 1–1 of 1 results for author: Ambegaokar, V

Estimating errors reliably in Monte Carlo simulations of the Ehrenfest model

Authors: Vinay Ambegaokar, Matthias Troyer

Abstract: Using the Ehrenfest urn model we illustrate the subtleties of error estimation in Monte Carlo simulations. We discuss how the smooth results of correlated sampling in Markov chains can fool one's perception of the accuracy of the data, and show (via numerical and analytical methods) how to obtain reliable error estimates from correlated samples. Using the Ehrenfest urn model we illustrate the subtleties of error estimation in Monte Carlo simulations. We discuss how the smooth results of correlated sampling in Markov chains can fool one's perception of the accuracy of the data, and show (via numerical and analytical methods) how to obtain reliable error estimates from correlated samples. △ Less

Submitted 4 June, 2009; originally announced June 2009.

Journal ref: Am.J.Phys.78:150-157,2010