The original structure-conduct-performance (SCP) paradigm, according to which market structure determines market conduct and market conduct determines market performance, underlies numerous competition policies. Since its development almost a century ago, the paradigm has been heavily criticised and numerous efforts have been made to test it by correlating measures of seller concentration with measures of market performance. The reliability of seller concentration measures that are frequently used, particularly in South Africa, was tested against the Hannah and Kay criteria, using hypothetical numbers of sellers and market shares. The premise is that a concentration measure must be reliable in the sense that it should lead to a correct conclusion when the relevant concentration curves do NOT cross. The following absolute concentration measures were found to meet the criteria: the Herfindahl-Hirschman index (HHI), the other Hannah and Kay indices [HKI(α)], the Rosenbluth index (RI), the numbers equivalent of the Hannah and Kay indices [HKIne(α)] and the entropy coefficient (EC). The discrete measures, concentration ratios (CRX) and the occupancy count (CRX%), do not always meet the criteria, nor do the relative concentration measures or measures of inequality, namely the Gini coefficient (GC), the variance of logarithms of market shares (VL) and the relative entropy coefficient (REC). The Horvath index (HI), an absolute concentration measure, does not always meet the criteria. Studies that employed the unreliable measures should be disregarded or reworked and students should be forewarned against the use of such measures.