The simplicity of the point-source stochastic simulation method makes it one of the most appealing tools for the quantification of ground motions for seismic hazard related studies. In this article, we scrutinize the limitations of this technique in terms of fundamental geophysical model parameters. To achieve this objective, we use the estimations of recent Next Generation Attenuation (NGA) and European empirical ground-motion models that are based on global strong-motion databases. The generated synthetics account for the local nonlinear soil effects through 1D site-response analysis. Thus, apart from our major objective, we also derive a probability-based soil-profile model that considers the random variation of shear-wave slowness as a function of depth. We discuss the critical variations of dynamic material properties, such as shear modulus and material damping, during the site-response analysis. We generate a total of 6000 synthetic records with a magnitude range of 5.0≤Mw≤7.5 and source-to-site distances (d) less than 100 km. The site class of the synthetics is defined by the average shear-wave velocity of the upper 30 m soil profile (VS30), with VS30 values ranging between 180 m/sec and 1500 m/sec. Our analysis indicates that synthetics that are generated using Rhyp (hypocentral distance) as the reference distance metric may fail to describe compatible variations of ground-motion demands with respect to global ground-motion prediction equations (GMPEs), particularly for small-magnitude and short-distance recordings. The performance of synthetics for these latter cases can be improved if one uses Rrup (shortest distance from the fault rupture). Our inspections on the frequency-domain behavior of synthetics reveal that they can be reliably used in spectral calculations for periods up to 20 sec. Within the context of this article, we also present a simple baseline correction method to obtain reliable ground displacements from the synthetics that are subjected to site-response analysis.