From its origins in Madagascar in the mid-1980s, the System of Rice Intensification (SRI) has generated considerable attention and investment in more than 30 countries. The primary logic supporting SRI promotion has been the contention that rice has untapped production potential that is not realized through conventional management practices like persistent flooding and use of inorganic fertilizers. SRI advocates maintain that synergies among its suite of simple management innovations unlocks the biological potential of rice and soil, and that this method also improves sustainability while being accessible to poor farmers, primarily because of a low reliance on external inputs. To date, much of the debate over the putative benefits of SRI has been theoretical or speculative. In aggregate, sufficient empirical data now exists to place SRI performance in a meaningful context by evaluating the productivity of SRI with respect to conventional best management practices (BMP). For this retrospective analysis, 43 site-years of SRI versus BMP comparisons were assembled into a common database. In addition to data from Madagascar, findings were compiled from Nepal, China, Thailand, Laos, India, Sri Lanka, Indonesia, Bangladesh, Philippines, and the Ivory Coast. Aside from one set of experiments in Madagascar where SRI more than doubled rice productivity with respect to BMP, we found no evidence of a systematic or even occasional yield advantage of this magnitude elsewhere. None of the 38 other experimental records demonstrated yield increases that exceeded BMP by more than 22%. Excluding the Madagascar examples, the typical SRI outcome was negative, with 27 of 38 site-years demonstrating inferior yields to best management and a mean deviation of -11%. With recognition that SRI yields in Madagascar are substantially below theoretical productivity levels, we find no evidence in the empirical record that SRI fundamentally changes the physiological yield potential of rice. On the other hand, there are reports from several countries that SRI substantially improves rice yields above those achieved with typical farmer practices. Is farmer practice rather than best management the only meaningful standard for comparison? If there are few unique biophysical synergies associated with SRI, we argue that efforts to enhance livelihoods in rice-based systems should take an undogmatic view of all management options with recognition that there are many plausible routes towards improved productivity. Indeed, there are facets of SRI that may make it ill-suited or impractical for many circumstances including: lack of sufficient organic composts to achieve reasonable yields over large areas, prevalence of hydric landunits were precision water control and drainage is impossible, and onerous labor requirements that retard sustained adoption. Moving forward, the rice research and agricultural development communities can take advantage of the renewed attention that SRI has brought to basic rice cultural practices to advance site-specific management approaches that are productive, efficient, and appropriate for farmers in marginal environments. Aspects of SRI may contribute to this effort, but after several years of attention and investment in over 30 countries, there is no evidence that justifies the promotion of SRI as a global model for rice production.