Computational experiments have been performed to study projectile erosion during the various phases of perforation of finite thickness steel plate targets. Studies were conducted at several impact velocities for thin, three-plate, spaced targets separated by one plate thickness. For tungsten alloy projectiles (Wi-Ni-Co), it is shown that the primary phase erosion is quantitatively similar irrespective of impact velocity and results in a perforation efficiency, η p , normally attained only in semi-infinite target, hypervelocity penetration cases. The indifference with respect to velocity is due to the close proximity of the plate rear surface and results in a similarity of the instantaneous perforation efficiency profile through the plate thickness. Further, it is shown that values for the semi-infinite penetration efficiency, as a function of impact velocity, vary during the penetration event and only attain their respective reference values at late-time. In short, P/L is not as velocity differentiated at early-time as it is at late-time. Bulge phase and residual phase erosion results are consistent with the notion that these erosion mechanisms are sensitive to impact velocity. For comparative purposes, simulations were also performed for a single-plate, equivalent line-of-sight (LOS) thickness plate target. In this case, the primary phase only perforation efficiency increases slightly with impact velocity. Taken together, perforation efficiency results for these two finite thickness plate targets and archival data for the penetration efficiency of semi-infinite thick targets suggest a small geometric effect when corresponding data are compared.