Enhancing microbial reduction of the potent greenhouse gas nitrous oxide (N2O) to N2 could be a promising strategy to mitigate emissions from soils, but N2O reduction rates are currently neither well understood nor quantified. It has been suggested that the importance of N2O reduction to N2 could be estimated from relationships between δ18O and δ15N when N2O production and reduction occur simultaneously. We assessed the robustness of such relationships by using simple simulation models and experimental data for various scenarios, including open versus closed system isotope dynamics and collection of N2O fluxes over time with and without refreshing the headspace between sampling. We found that relationships between δ18O and δ15N vary dramatically with experimental conditions (such as headspace gas concentration and sampling scheme) and open versus closed system isotope dynamics, irrespective of N2O reduction rate. Therefore, we conclude that the simple relationships between δ18O and δ15N currently used are not robust indicators of N2O reduction to N2 and strongly discourage graphical interpretations of δ18O and δ15N to determine N2O reduction. We recommend the development of more advanced process-based isotope models that take into account open versus closed system isotope dynamics, experimental conditions, isotope values of precursors and other biogeochemical controls on δ15N and δ18O of N2O to estimate N2O reduction rates more reliably from natural abundance isotope values of N2O.