In this study, a total of 191 cases with STR exclusions out of 591 paternity cases were analysed using 2 STR sets, i.e. (set a) 5 STRs in 462 cases with 150 exclusions and (set b) 9 STRs in 129 cases with 41 exclusions. Set (a) was associated with four exclusions on average while set (b) showed five exclusionary loci on average. Double exclusions were observed in 18 cases and further elaborated. Of these, 2 ended up with probabilities of paternity of 0.1% and 0.4%, respectively and with a random occurrence of the hypothesis "mutation" of 1:20,000 and 1:50,000, respectively, while all other cases were associated with much lower frequencies. The conclusion is that the evidential value of a set of highly polymorphic STRs applied in paternity cases is usually extremely high.