In one embodiment, an initial series of bytes are scanned to determine a frequency of occurrence within the initial series of bytes for a plurality of byte strings. A data structure is generated containing the plurality of byte strings and count values representing a frequency of occurrence for the plurality of byte strings in the initial series of bytes. A most common byte string of the plurality of byte strings in the initial series of bytes is determined from the count values. Each instance of the most common byte string in the initial series of bytes is replaced with a first substitute byte string to form a revised series of bytes. A most common byte string in the revised series of bytes is determined from the data structure without resetting the count values associated with at least some of the plurality of byte strings.