The output of a simple statistical categorizer is used to improve recognition performance on a homogeneous data set. An array of initial weights contains a coarse description of the various classes; as the system cycles through a set of characters from the same source (a typewritten or printed page), the weights are modified to correspond more closely with the observed distributions. The true identities of the characters remain inaccessible throughout the training cycle. This experimental study of the effect of the various parameters in the algorithm is based on ~30 000 characters from fourteen different font styles. A fivefold average decrease over the initial rates is obtained in both errors and rejects. © 1966, IEEE. All rights reserved.