Most of the systems were purposely defensive in nature. One version of realist theory is offense-defense realism. Defensive realism holds that when the strength of states is defensive in nature it will mitigate against interstate conflict because security is maximized. Conversely when offensive strength is the norm, it is more likely that states will engage in conflict with one another in an attempt to maximize their own security.
Both liberalism and realism in most of their varieties assume that the main unit of analysis in international relations should be the state. The difference between them is over the concern for power (absolute gains) versus the concern for economic prosperity (relative gains) (Walt 1998, 38). In Neorealism and Neoliberalism, The Contemporary Debate (Baldwin, Ed. 1993) it is shown as evident that the only true difference between the two main strands of liberal and realist theory is the question of when absolute gains mean more than relative gains. .
For the student of international relations theory and for the practitioner looking for a way to provide the best foreign policy this becomes quite troubling. Theories are supposed to provide viable options that are competitive in their ambition to explain a phenomenon or to provide a means of predicting future behavior and outcomes. This is a general theoretical problem, but as Hans Morgenthau has informed us, "The key to a theory of international politics will not be found in the specific subject matter of international politics but in the requirements and problems of a general political theory," (1959, 16). .
Stephen Van Evera says that theory is composed of causal law or causal hypothesis connected with an explanation of the causal law or hypothesis (1997, 9). The implication of this definition for social science is that theory should allow us to make predictions about future behavior of the actors involved based on an understood explanation of how phenomena relate to each other.