In classical occupancy receptor theory, efficacy is a dimensionless proportionality constant denoting the power of agonists to produce a pharmacological response. In theoretical terms, it is difficult to separate affinity and efficacy estimates of agonists for receptors, hence questioning the value of clearly flawed estimates of efficacy by conventional methods. In this paper, the use of efficacy estimates, the limitations of the current methods to estimate efficacy, and the types of systems in which serious errors in efficacy estimation would be expected, is discussed. Specifically, in constitutively active receptor systems or in those where the receptor interacts with more than one G protein, there are theoretical objections to the use of relative maximal responses as indicators of intrinsic efficacy of agonists.