This paper makes two methodological contributions to the sparse literature on replication studies in quantitative finance and accounting research. First, it demonstrates the importance of accurate and complete disclosure of sample reductions in quantitative empirical accounting and finance research. The sample reduction should perfectly reconcile with the sample selection, i.e., the number of observations in descriptive statistics and result tables. Second, it proposes a formal framework to evaluate original against replicated results in the context of complex studies with multiple focal points, which are partial results on which the original authors focus to make their case. The study of Azevedo, Bielstein, & Gerhart (2021) serves as a showcase. My replications show that about half of the results which Azevedo et al. provide in favor of their new combined model (CM) hold. Raw analysts' forecasts are more accurate and not more biased than forecasts based on the CM.