Although there is already a wealth of literature that examines the causes and impacts of war on the international system, there has been little analysis of the decisions of leaders who initiate wars they are destined to lose. The failures of these aggressors have been variously explained as the inevitable occurence of "balancing behaviour" of international alliances, as over eagerness on the part of a would-be powers, or as a product of the inherent riskiness of war, but all are found wanting. Balances often do not arise, and when they do they are frequently the creation of the war initiator. As to cycles of rising and declining power, this study finds little evidence of losing wars begun by the impatient. With regard to chance, the "loser" phenomenon's noticeable patterns suggest that there is more to it than the rolling of a dice. The insight that this study introduces is that all interstate wars are not created equal. There are two fundamental types, requiring different analytical treatment; land wars and land-sea wars. Leaders from Louis XIV to Saddam Hussein have, at best, an imperfect understanding of land-sea wars, and the author suggests that this is at the root of the defeats they have eventually suffered. Both quantitative and case study methods are applied in the testing of this "dual" theory of war. The case studies serve also to illuminate the finer points of the theory, and to illustrate Clausewitz's thesis that "war is a continuation of policy by other means," from which the work begins. The study concludes by considering some implication of its findings for organizations, policy, and deterence. It is aimed at Government policymakers, non-government officials, professors and graduate students in international relations, international affairs, foreign policy, security studies and world history.