Discusses a new error concept dispensing with the common practice to randomize unknown systematic errors. Instead, unknown systematic errors are treated as what they physically are - namely as constants being unknown with respect to magnitude and sign.