The tacit assumption in early severe accident studies was that the melting of a reactor core would result in failure of the reactor pressure vessel and eventual failure of the containment building and release of fission products to the environment. This assumption was shown to be wrong by the Three Mile Island Unit 2 (TMI-2) accident in which 50% of the core melted, yet fission product release to the environment was insignificant (<5% of the noble gases). Termination of the accident and survival of the reactor vessel is attributed to the presence of water in the vessel. The quantity of water required to cool the molten core and reestablish stable core cooling is calculated. These calculations and the TMI-2 experience imply that future accident management strategies that emphasize restoration of the cooling water supply can terminate advanced severe accidents and avoid pressure vessel failure.