Numerical methods used to solve ordinary and partial differential equations fall into a few basic categories. However, no traditional PDE method is well suited for solving systems of equations which have highly irregular grids. For many problems involving high velocity material impact or galaxy-galaxy collisions, tangled grids are inevitable. Since the computational cost of regridding can be considerable, we have turned to statistical techniques which work well with highly irregular data. In this talk, we present a historical overview of PDE and ODE solvers. We will also present an overview of Smoothed Particle Hydrodynamics (SPH) and show its connection to kernel density estimators used in statistics.