The Paris Peace Conference was arguably the most complex negotiation ever undertaken. The principal product of the conference, the Treaty of Versailles, failed to accomplish any of its framers' major goals. Relations between the Allies themselves and the Allies and their defeated enemies seriously deteriorated as a consequence of the negotiations and attempts to implement the treaty. Economic conditions in Germany, the rest of Europe, and eventually the United States declined as well. At the time of the Treaty's publication, John Maynard Keynes and a considerable number of other participants predicted these events, pointing to the negotiators' errors and oversights as a primary cause. The logic of Keynes' argument is re-examined in light of recent research on the psychology of human information processing, judgment and choice. It reveals that his approach is actually very consistent with and anticipates both Simon's conception of bounded rationality and recent work on cognitive heuristics and illusions. Negotiator bias has been studied almost exclusively using simple laboratory settings. The catastrophic lose-lose nature of the Versailles Treaty illustrates the way in which complexity necessitates reliance on simplifying heuristics while propagating and amplifying the impact of the bias that is generated. Evidence from the treaty negotiations and the failed implementation of the treaty suggest some very significant boundary conditions for the application of rational choice models in the business, politics, and international relations contexts. It also demonstrates the need for negotiations researchers to focus more attention on the implementation of agreements and the long-term effects of those agreements on relationships.