The paper presents an alternative approach to constructing point charge water models, which are widely used in atomistic molecular simulations, particularly in biology and chemistry. The authors propose a method that optimizes the distribution of point charges to best describe the electrostatics of the water molecule, rather than imposing geometric constraints. This approach is based on optimizing the lowest multipole moments of the water model to match experimental bulk water properties at room temperature. The resulting 3-charge, 4-point rigid water model (OPC) reproduces a comprehensive set of bulk water properties with an average error of 0.76% relative to experiment, and shows close agreement over a wide range of temperatures. The OPC model also predicts hydration free energies of small molecules more accurately than commonly used rigid models, with a root-mean-square error of less than 1 kcal/mol. The improved accuracy of the OPC model is attributed to its focus on electrostatics and the unrestricted fine-grain search in the multipole moment space. The authors suggest that this approach can be extended to develop more accurate water models with different numbers of point charges and flexible or polarizable models.The paper presents an alternative approach to constructing point charge water models, which are widely used in atomistic molecular simulations, particularly in biology and chemistry. The authors propose a method that optimizes the distribution of point charges to best describe the electrostatics of the water molecule, rather than imposing geometric constraints. This approach is based on optimizing the lowest multipole moments of the water model to match experimental bulk water properties at room temperature. The resulting 3-charge, 4-point rigid water model (OPC) reproduces a comprehensive set of bulk water properties with an average error of 0.76% relative to experiment, and shows close agreement over a wide range of temperatures. The OPC model also predicts hydration free energies of small molecules more accurately than commonly used rigid models, with a root-mean-square error of less than 1 kcal/mol. The improved accuracy of the OPC model is attributed to its focus on electrostatics and the unrestricted fine-grain search in the multipole moment space. The authors suggest that this approach can be extended to develop more accurate water models with different numbers of point charges and flexible or polarizable models.