Experiences with the multi-level algorithm

Small expectation values are difficult to measure in Monte Carlo calculations as they tend to get swamped by noise. Recently an algorithm has been proposed by Luscher and Weisz which allows one to measure expectation values which previously could not be measured reliably in Monte Carlo simulations. We will test our implementation of this algorithm by looking at Polyakov loop correlators and then explore ways of applying it for measuring large Wilson loops.