Multi-body simulations of a wide range of impact conditions show that the distance a pedestrian is displaced in a crash is inherently sensitive to impact parameters such as relative pedestrian-vehicle geometry, pedestrian pre-impact orientation, and ground frictional characteristics. Thus, reconstruction techniques, such as the use of throw formulae that solely rely on pedestrian throw distance for vehicle impact velocity prediction have large uncertainty associated with them. However, using crash simulators, such as PC-Crash, or more advanced multi-body modeling programs, such as MADYMO, to account for other crash variables requires iterative simulations at considerably increased modeling time. This study compares the absolute accuracy of three methods of impact velocity prediction as well as their relative accuracy as a function of computational time.