For three decades, molecular models for water, nature＇s most important liquid, have been developed and refined by fitting to structure measured by neutron scattering. The decade-old widely accepted structure of water at room temperature and pressure was recently revised as a byproduct of attempts to understand the structure of high-temperature/high-pressure water for which, in a remarkable reversal of roles, molecular models successfully pinpointed inaccuracies in scattering data. Subsequent improvements in analyzing scattering data have led to reevaluation of water structure at normal conditions. This remarkable interplay between molecular modeling and experiment suggests molecular methods can effectively complement scattering experiments.