Abstract Objective To compare the predictive ability for oncologic outcomes among current tumor size cut-points and clinically relevant alternatives in order to determine which are optimal. Methods Patients who underwent radical or partial nephrectomy between 1970-2010 for T1-2Nx/N0M0 renal cell carcinoma (RCC) were identified. Associations between tumor size and progression-free survival (PFS) and cancer-specific survival (CSS) were evaluated using Kaplan-Meier analyses and Cox models. Predictive ability was assessed using c-indexes. Results The cohort included 3304 patients with a median age of 63 years (IQR 53,70). Median follow-up among survivors was 9.9 years (IQR 6.9,14.3). There were 536 patients who progressed and who 354 died from RCC. For RCC tumors ≤3.0cm, 10-year PFS and CSS rates were 93-95% and 97-99%, respectively. For tumors >3.0-4.0cm, PFS and CSS began to decline (91% and 95%, respectively), with further gradual declines in PFS and CSS with increasing tumor size. Plots of hazard ratios for progression and RCC-death as a function of tumor size did not reveal major inflection points. Differences in discrimination based on various combinations of tumor-size cut-points for progression or RCC-death were small, with c-indexes ranging between 0.691-0.704 and 0.734-0.750, respectively. Conclusions RCC tumors ≤3.0cm in size are associated with favorable outcomes. Thereafter, risks of progression and RCC-death increase gradually with tumor size, with no compelling biological reason to endorse a given cut-point over another.