Outliers are very common in noisy data. Unfortunately outliers can bias estimates of the moments. In this case am interested in the covariance of a data set. In addition to being prominent in stochastic equations, portfolio analysis (the usual stuff in finance), it is also used in linear projection and prediction.