Skip to Main Content
The paper describes the application of some statistical estimation techniques to the problem of evaluating the probability of error of (digital) communication systems. The usefulness of an estimator is governed by the variance of the estimate, which, in this case, is the measured probability of error. It is shown that, for the low values of probability of error considered to be practically useful, simple sample-mean estimation requires somewhat large numbers of statistical trials to give results of low variance. The number of statistical trials, however, may be decreased by the use of simple control or antithetic estimators.