Breast cancers are routinely assessed for estrogen receptor status using immunohistochemical assays to assist in patient prognosis and clinical management. Specific assays vary between laboratories, and several antibodies have been validated and recommended for clinical use. As numerous factors can influence assay performance, many laboratories have opted for ready-to-use assays using automated stainers to improve reproducibility and consistency. Three commonly used autostainer vendors-Dako, Leica, and Ventana-all offer such estrogen receptor assays; however, they have never been directly compared. Here, we present a systematic comparison of three platform-specific estrogen receptor ready-to-use assays using a retrospective, tamoxifen-treated, breast cancer cohort from patients who were treated in Calgary, Alberta, Canada from 1985 to 2000. We found all assays showed good intra-observer agreement. Inter-observer pathological scoring showed some variability: Ventana had the strongest agreement followed closely by Dako, whereas Leica only showed substantial agreement. We also analyzed each estrogen receptor assay with respect to 5-year disease-free survival, and found that all performed similarly in univariate and multivariate models. Determination of measures of test performance found that the Leica assay had a lower negative predictive value than Dako or Ventana, compared with the original ligand-binding assay, while other measures-sensitivity, specificity, positive predictive value, and accuracy-were comparable between the three ready-to-use assays. When comparing against disease-free survival, the difference in negative predictive value between the vendor assays were not as extreme, but Dako and Ventana still performed slightly better than Leica. Despite some discordance, we found that all ready-to-use assays were comparable with or superior to the ligand-binding assay, endorsing their continued use. Our analysis also allowed for exploration of estrogen receptor-negative, progesterone receptor-positive cases, and we discovered that this phenotype was not consistent across the assays, suggesting this might be an artifact.