The key to extended deterrence, at least under parity, lies in the way threat credibility is arrayed. This much is clear from our discussion of the Asymmetric Escalation Game with complete information. But many important questions about extended deterrence remain unanswered: in an uncertain world, how should multi-level threats be fashioned to best ensure the survival of the status quo? What is the best way to connect lower-level (or tactical) threats and higher-level (or strategic) threats? Are limited wars possible between two equally capable states and, if so, can threats be configured to ensure that they remain limited? Under what conditions are extended deterrence relationships likely to spiral to the highest level and culminate in an all-out conflict?

Historically, two broad schools of thought exist with respect to these and related questions: the all-or-nothing approach and the limited-war approach. In an insightful discussion of post-war American defense policy, Gacek (1994) traces the origins of the all-or-nothing school to a set of strategic principles enumerated by the Swiss military writer Henri Jomini. Drawing on Napoleon's legacy, Jomini argued that success in battle came from the decisive application of superior force with the purpose of confronting and exploiting an opponent's greatest vulnerability.