Single-letter statistical measures providing values for each letter-position and word-length combination are too unwieldy for use in working letter-recognition models; more compact tables are needed. Compact tables collapsing word length to short or long words and letter positions to first, last, or middle letters are presented for the frequency and versatility of single letters. Letter-position and word-size differences are preserved in this reduced format. To test awareness of these values, subjects rate the commonness of letters in each letter position. Their responses indicate high intersubject agreement and correlate highly with the frequency and versatility measures. A LISP program that translates the letter knowledge for each letter into a corresponding knowledge for each feature in a feature set is described. Distinctiveness values for each feature (see Shimron & Navon, 1981) are computed.