In recent literature the gradual yielding of quenched martensitic steels has been attributed to either heterogeneous microstructures having different strength levels or to the presence of small scale, Type II, residual stresses. Using in-situ tensile testing in synchrotron diffraction experiments in combination with crystal plasticity finite element modelling (CPFEM) we show that the dominant influence on yielding derives from the residual stresses which are a product of the displacive transformation from austenite during quenching. As plastic straining proceeds, the measured diffraction peaks become narrower and asymmetric, as predicted by the CPFEM calculations. The model predictions are generally in good agreement with published results showing large variations in local strains in different microstructural elements.