In political science, there is an on-going debate about the usefulness of rational choice theory in analyzing particular real world situations. The rational choice school has been known for its ability to develop general theories. With its methodological sophistication, it has developed prominent theories with many testable hypotheses, but little attention was paid to the individual and real-world political events by this school. Critics of rational choice point to the abstract and logical character of game theory, but do so in order to condemn it. The failures of rational choice theory, Green and Shapiro (1994) declare, are rooted in the aspiration of rational choice theorists to come up with universal theories of politics. The result, they argue, is a preoccupation with theory development, accompanied by a striking paucity of empirical application (1994, x). Research becomes theory driven rather than problem driven; its purpose is to save or vindicate some variant of rational choice theory, rather than to account for political phenomena (1994, 6). Recently Bates et al. (1998) introduced the notion of analytic narratives. According to them, analytic narrative combines analytic tools that are commonly employed in economics and political science with the narrative form, which is more commonly employed in history. It is narrative; it pays close attention to stories, accounts, and context. It is analytic in that it extracts explicit and formal lines of reasoning, which facilitate both exposition and explanation. It is problem driven, not theory driven; it is motivated by a desire to account for particular events or outcomes. It is devoted to the exploration of a case, not to the elaboration of theory. In these ways, it counters the charges raised by Green and Shapiro. Although informed by deductive reasoning, it seeks no universal laws of human behavior (1998). In this paper, I employ the analytic narratives approach to account for a particular event. The real world event of interest is the changing relationship between North and South Korea and one between the U.S. and North Korea, and the methodological tool applied is game theory. As I will demonstrate in the paper, both South Korean and U.S. policy makers have attempted to formulate their new North Korea policies while they were uncertain about the true intentions of the North Korean leadership and its weapons capabilities, especially since the late 90s. By developing two Bayesian models of the Korean situation, I make some very interesting predictions and policy recommendations. One potential payoff of developing game-theoretic models of a real world situation such as that on the Korean peninsula is that I can make predictions and policy recommendations specific to the situation and observe empirically how the players act as the situation unfolds. [ABSTRACT FROM AUTHOR]