Finite Element simulations have become a popular tool to investigate the mechanisms of launch dynamics. However, the common techniques for evaluating the simulation results in terms of target impact accuracy are far from ideal. This paper presents a method for the postprocessing of such simulations. The approach is based on a decomposition of the bullet's transverse motion during launch into a regular rotation caused by the rifling, irregular oscillations (balloting) and muzzle motion. Target impact dispersion is evaluated separately for each of these motion components, which provides a deeper understanding of the dynamic processes determining the weapon's hit performance. In contrast to traditional methods, where the assessment of dispersion requires large numbers of simulations with varying launch conditions, the evaluation of accuracy is based on a single simulation.