The author presents the theory of portfolio choice from a new perspective, recommending decision rules that have advantages over those currently used in theory and practice. Portfolio choice theory relies on expected values. Goodall argues that this dependence has a historical basis and argues that current decision rules are inadequate for most portfolio choice situations. Drawing on econometric solutions proposed for the problem of forecasting outcomes of a chance experiment, the author defines adequacy criteria, and proposes adequate decision rules for a variety of situations. Goodall's theory combines the problems of prediction and choice, and formulates solutions based on cost functions that fit the underlying decision situation.