The performance of sub-slab-ventilation (SSV) systems has been parametrically studied with a numerical model that was earlier compared successfully with experiment (Bonnefous et al., 1992). The model distinguishes between the sub-slab gravel and the underlying soil. It is used w examine system performance for the following system parameters: the permeability of the soil and of the sub-slab gravel, the magnitude of pressurization (or depressurization) applied by the SSV system, and the mode of SSV application (i.e. pressurization (SSP) or depressurization (SSD)). The mechanisms contributing to the successful performance of SSP and SSD systems are identified. For SSD systems, the mechanisms are (1) the inversion of the pressure gradient across the basement slab, and (2) the reduction of the radon concentration in the soil. For SSP systems, the mechanisms are (1) the elimination of convective flow of soil-gas from the soil into the sub-slab gravel by pressurization of the sub-slab region, (2) the reduction of the radon concentration in the soil, and (3) the suppression of diffusion of soil-gas from the soil into the sub-slab gravel by advective flow of air from the gravel bed into the subgravel soil. Numerical modeling demonstrates that placement of a sub-slab gravel layer substantially improves the SSV system performance. Except in the case of highly permeable soils, SSD systems are predicted to perform better than SSP systems. This prediction is consistent with reported field experience. The numerical model is used to elucidate the reasons for this difference in performance.