<p>This study examines the role of shear stiffness and crack initiation thresholds in predicting the maximum extent of failure around excavations in sedimentary rocks. Data from 558 unconfined compressive strength tests were analysed using the volumetric strain response method to determine the crack initiation thresholds for sandstone, siltstone, and coal. The findings provide robust crack initiation thresholds for sedimentary rocks based on a dataset that contains 100 times more sedimentary rock results than those currently available in existing databases. A new failure criterion for estimating the extent of failure, the transversely isotropic brittle enhanced representation (TIBER), has been developed to account for shear stiffness stress dependency that eliminates the assumptions required in existing models. Case studies with modelling predictions using an isotropic criterion, the transversely isotropic brittle (TIB) criterion, and the newly proposed TIBER criterion are compared with in situ observations. The results show that isotropic models significantly underestimate the extent of failure. Using the TIB criterion, with constant shear stiffness, only marginally improves the predictions. In contrast, using the TIBER criterion, with shear stiffness implemented as a function of effective normal stress, produces excellent extent of failure predictions. This study emphasises the necessity of implementing shear stiffness as a function of applied effective normal stress to model the transversely isotropic response of bedded rock masses adequately. This novel approach significantly improves the prediction of the maximum extent of failure in sedimentary rocks, providing a more effective and sustainable method for roof support design.</p>