Despite being accepted as a robust assessment of masonry stability, thrust line analysis (TLA) relies on assumptions that can lead to a conservative assessment of stability. This article aims to quantify the extent of these limitations through a comparison of TLA with discrete element modeling (DEM). Two studies are provided. The first study compares TLA with DEM (using fixed input parameters) in assessing the stability of unreinforced masonry arches, semi-circular barrel
vaults, multi-ring arches, and groin vaults. The tests demonstrate the types of sliding failures overlooked by the safe theorem due to its assumption of infinite friction. Following these validations, the comparisons between 2D structures and 3D counterparts also give insight into the efficacy of the slicing method. The second study examines the effect of DEM input parameters on the DEM-predicted stability of the considered geometries. While material parameters had limited effect on the determination of stability, for each typology, joint friction angle had a unique impact on stability. These trends are graphically presented and demonstrate how t/R ratios alone are not sufficient to unequivocally confirm stability of the considered vaults. Overall, this research informs the extent of safety for using the geometry-based analysis tool, TLA, for
analyzing masonry structures.