In this report, the authors describe an experimental "thinking-Red" approach to analysis, wargaming, and other exercises that might help inform strategies to avoid aggression or escalation in a crisis. The approach can be applied to (1) diverse potential adversaries, (2) direct and gray-zone conflicts, and (3) different levels of crisis, conflict or competition. The approach features alternative models of the adversary.
- What strategies might help avoid aggression or escalation in a crisis?
- What mix of deterrence (by threat of punishment), denial, dissuasion by any means, reassurances, and incentives would create a successful influence strategy?
- What political, military, and economic instruments of power might produce a successful strategy?
- How are the answers to these questions affected by recognizing our uncertainty about adversary perceptions, motivations and capabilities?
In this report, the authors describe an experimental "thinking-Red" approach to analysis, wargaming, and other exercises that may help inform strategies to avoid aggression or escalation in a crisis. This thinking-Red approach focuses on how to influence an adversary's reasoning in ways that the decisionmaker regards as favorable. It does so with alternative models of the adversary. An influence strategy might involve a mix of deterrence by threat of punishment, deterrence by denial, dissuasion by many means, reassurances, and incentives. Deterrent threats alone will seldom constitute effective strategy and, depending on the adversary's motivations and perceptions, could even be counterproductive. A successful strategy will also often require artful orchestration of political, military, and economic instruments of power.
The approach can be applied to (1) diverse potential adversaries, (2) both direct and gray-zone conflicts, and (3) different levels of crisis, conflict or competition. Each of the three applications will require an in-depth study of substantive issues and refining methods and tools, but the potential scope of applications is wide.
Deterrent efforts should be seen as elements of much broader influence strategies
- Deterring adversaries by threat of punishment is often ineffective by itself. Better influence strategies include, e.g., being able to thwart aggression, reassurances to reduce the adversary's fears, incentives, conditional reciprocity, relationship-building, and both direct and indirect forms of dissuasion.
- Influence strategies depend on affecting Red's thinking, not just displaying strength.
- Considering alternative models of Red can improve strategy and its adaptiveness.
Understanding how an adversary reasons and perceives U.S. actions will help decisionmakers develop effective influence strategies
- Influence strategies should be guided by thinking Red, but this is notoriously difficult. It is very helpful to identify alternative ways that Red may be thinking, to take these alternatives seriously, and to pursue an adaptive strategy likely to be effective for all of them.
- A decision-aiding doctrine for doing so would be quite different from the common approach of focusing on a best-estimate image of Red, which often proves wrong.
- A method called uncertainty sensitive cognitive modeling could help with effective thinking-Red and could suggest elements of influence strategy.
- Such modeling could include gaming, other human exercises, and qualitative modeling to create a coherent depiction of insights.
- Prototype experiments suggest that such a process is feasible and could be insightful. The approach suggested is ready for more-realistic testing in a variety of context.
- Embrace uncertainty sensitive cognitive modeling and apply it experimentally to a variety of real-world efforts to influence great-power adversaries in ways that are favorable to the United States. These applications could involve heading off direct aggression, heading off aggression achieved through gray-zone activities, different levels of conflict, and anticipating competitor actions against allies and partnerships.
Table of Contents
Lessons from History and Psychology
Cognitive Modeling and Thinking-Red Exercises
Illustration: Hypothetical Crisis with China
Mathematics of Thresholded Linear Weighted Sums
Computational Version of Factor Tree Model
Bayesian Updating of Red's Blue: Simple Artificial Intelligence
Input Data for Model
Simplifying Probabilistic Calculations