A conventional assumption in economic models is that expert judgement requires Bayesian behavior. A justification for this assumption is that because Bayesian behavior results in superior decisions, it is dominant in a evolutionary sense. Bayesian experts who are sequentially rational, however, set standards that are "too low". Consequently, heuristic behaviors that result in more stringent sequentially rational standards may dominate Bayesian behavior. This suggests that deviations from Bayesian behavior observed in experimental auditing settings may be viable in an evolutionary sense in real economic settings.