Active thrust faults pose a significant seismic hazard worldwide. Many of these faults include “blind” thrusts, where the propagating fault tip does not reach the Earth's surface, and “buried” faults, where the geomorphic expression of the fault is obscured by subsequent sedimentation and/or erosion. This raises the issue of whether conventional geologic, geomorphic, and paleoseismic methods can be used to identify and characterize thrust faults for the assessment of seismic hazards or whether these faults sometimes are truly “blind.” We compiled a data base of 148 worldwide moderate- to large-magnitude thrust/reverse earthquakes to evaluate whether or not the event occurred on a fault that could have been identified prior to the earthquake on the basis of recognizable Quaternary surface deformation (i.e., a pre-existing fault or fold). Analysis of the data shows that interplate reverse earthquakes almost always are associated with pre-existing Quaternary deformation that was or could have been recognized prior to the earthquake. In particular, most interplate reverse earthquakes are associated with an active reverse fault at the surface and/or an active anticline. In contrast, intraplate reverse earthquakes seldom occur on faults associated with pre-existing recognizable surface deformation. We conclude that thrust faults can be detected in interplate regions with careful Quaternary geologic and geomorphic mapping; furthermore, the absence of Quaternary surface deformation can be used to infer the absence of an underlying active blind thrust fault in interplate tectonic settings. However, the data show that Quaternary geologic mapping techniques alone likely are insufficient to characterize blind thrusts in intraplate regions. In these areas, inclusion of a floating or random earthquake may be necessary to assess earthquake hazards.