Structure‐activity relationships (SARs) are the principle resource for estimating rate coefficients that have not been measured experimentally. With the advent of large, near‐explicit oxidation schemes, especially those that are automatically generated, 1,2 the need for estimation techniques has increased dramatically. Notwithstanding, few studies are available that assess the applicability, accuracy and predictiveness of SAR estimates. In this study, we estimate organic compound oxidation rate coefficients by hydroxyl, ozone, nitrate and chlorine. For this purpose, we have compiled a large database that contains kinetic data for over 1100 compounds. By combining this large sample size together with a diverse range of estimation methods, we present a rigorous test of SAR performance that affords several insights: notably, which functionalities are poorly predicted, which species would benefit most from additional laboratory measurements, and whether an ensemble approach can yield a more robust prediction of chemical reactivity.