A method, system, and computer program product for segmenting data into variable size blocks based on content defined positions. Segmenting probabilities and associated segmenting conditions are defined. The segmenting conditions are ordered in accordance with the associated segmenting probabilities to form a hierarchy of the segmenting conditions. A segmenting condition associated with a highest segmenting probability is defined to be a lowest level segmenting condition in the hierarchy of the segmenting conditions. The segmenting condition associated with a lowest segmenting probability is defined to be a highest level segmenting condition in the hierarchy of the segmenting conditions. Hash values are calculated for each seed block in each consecutive byte position in the data. Each one of the hash values is evaluated using the segmenting conditions. A segmenting position is determined in the data for each hash value that satisfies one of the segmenting conditions.