The objective of this paper is to apply field data, without modification, for a number of case studies in different numerical models to find the best non-calibrated approach. The purpose of the comparison is to identify which of a number of well-known and frequently used approaches to failure prediction, that gives the best agreement with the observed fallouts with respect to location, depth, shape, and extent. Four case studies are presented which all are situated in hard rock masses that are massive or sparsely fractured, at depths of 630-965 m below surface. The observed fallouts are compared with the results from five numerical models in Examine2D and Phase2D by studying (i) strength factors (SF) and yielded elements and (ii) volumetric and (iii) maximum shear strains. This study demonstrates that when the rock mass strength parameters are defined by either the Hoek-Brown or the Mohr-Coulomb criterion the predicted fallout depth by the SF and yielded elements resulted in a reasonable agreement with the observed fallout depth. Still the extent on the boundary and shape is incorrect. As a preliminary result, the volumetric strain is not found to be a good indicator of fallouts of hard (high strength) rock masses. Also, due to the high rock mass strength, no shear bands were formed when evaluating the maximum shear strain, except when using instantaneous softening of friction and/or cohesion. However the instantaneous softening approaches significantly overestimated the predicted fallouts. Despite the fact that none of the approaches in this study showed perfect agreement with the actual fallouts, a 2D stress analysis program can be used for guidance of where, on the boundary, the actual compressive fallout will occur.