BACKGROUND: Log odds of positive lymph nodes (LODDS) is defined as the log of the ratio between the probability of being a positive lymph nodes and the probability of being a negative lymph nodes when one lymph node is retrieved. The value of LODDS staging system on prognostic assessment for gastric cancer patients with R0 resection is still unclear. METHODS: Clinicopathologic and prognostic data of 2547 gastric cancer patients underwent D2 or D3 lymphadenectomy with R0 surgery were retrospectively studied. RESULTS: Multivariate analysis indentified LODDS stage was an independent prognostic factor, but not pN classification or rN classification. The scatter plots of the relationship between LODDS and the number, the ratio of nodes metastasis, suggested that the LODDS stage had power to divide patients with the same number or ratio of nodes metastasis into different groups. For patients in each of the pN or rN classifications, significant differences in survival could always be observed among patients in different LODDS stages. However, for patients in each LODDS stage, prognosis was highly homologous between those in different pN or rN classifications. A minimum number of 10, 15, 20, 25, and 10 nodes retrieved should be met for patients in the pN0, pN1, pN2, pN3, and rN0-3 classifications, respectively, unless the hazard risks of death would be underestimated or overestimated. However, LODDS stage could discriminate among 5 groups of patients with highly homologous prognosis, regardless how many nodes retrieved. CONCLUSIONS: The LODDS system is more reliable than the Union Internationale Contre le Cancer and American Joint Committee on cancer pN system and the rN system for prognostic assessment.