I think the granular pile problem is still unsolved: Can you predict the cone angle of a pile of granular material given a small set of initial parameters?