He notes that fingers are different with respect to ridge detail, that minutiae in fingerprints are distributed randomly, and that fingerprint identification is similar to DNA identification. He explores the consequences of random distribution of fingerprint details, points out uncertainties associated with analyzing fingerprint ridge details, and examines probability models used to identify these details. Finally, the author recognizes that latent fingerprints are often distorted and contain a variety of defects and that a fingerprint identification entails more than just point and probability counting. He questions the ability of the CTS latent fingerprint proficiency test to absolutely establish a person's identity and suggests sensible consideration of probability issues when comparing fingerprints.