Abstract Background This study sought to develop a clinical risk score for resectable colorectal liver metastasis (CRLM) by combining clinicopathological and clinically available biological indicators, including KRAS. Methods A cohort of patients who underwent resection for CRLM at the Johns Hopkins Hospital (JHH) was analysed to identify independent predictors of overall survival (OS) that can be assessed before operation; these factors were combined into the Genetic And Morphological Evaluation (GAME) score. The score was compared with the current standard (Fong score) and validated in an external cohort of patients from the Memorial Sloan Kettering Cancer Center (MSKCC). Results Six preoperative predictors of worse OS were identified on multivariable Cox regression analysis in the JHH cohort (502 patients). The GAME score was calculated by allocating points to each patient according to the presence of these predictive factors: KRAS-mutated tumours (1 point); carcinoembryonic antigen level 20 ng/ml or more (1 point), primary tumour lymph node metastasis (1 point); Tumour Burden Score between 3 and 8 (1 point) or 9 and over (2 points); and extrahepatic disease (2 points). The high-risk group in the JHH cohort (GAME score at least 4 points) had a 5-year OS rate of 11 per cent, compared with 73·4 per cent for those in the low-risk group (score 0–1 point). Importantly, in cohorts from both the JHH and MSKCC (747 patients), the discriminatory capacity of the GAME score was superior to that of the Fong score, as demonstrated by the C-index and the Akaike information criterion. Conclusion The GAME score is a preoperative prognostic tool that can be used to inform treatment selection.