Can red teaming and alternative analysis help decision makers anticipate events? As you might expect, the answer is not a simple “yes” or “no.”
First, anticipating events is extremely difficult, and most analysts–futurists included–will tend to caveat their work heavily. This is partly due to the limitations inherent in every analytical technique and partly due to the analyst’s natural reluctance to leap at long shots. While this can lead to muffled insights and recommendations, it also helps avoid spectacular mistakes.
Qualitative analysis is often best suited to address human-driven complexity, but qualitative approaches generally sacrifice analytical precision for flexibility and descriptive power. Scenarios in particular are susceptible to problems related to the conjunction fallacy, and a persuasive narrative can actually lead a decision maker to misperceive risk.1 Quantitative analysis, on the other hand, gains in precision but often obscures its workings behind a curtain of formulas and algorithms, in many cases reducing a decision maker’s willingness to trust the results.
Second, analysis is not an all-or-nothing proposition. An analyst, for example, might predict one element of an event but fail to anticipate others, and the value of the anticipated element is certain to vary by context. Further, getting a foundational element wrong will often preclude getting a dependent element right.
The following model defines the various elements of anticipation and their relationships. It’s a rough, back-of-the envelope sketch, but it helps clarify the challenge.
- Strategic elements form the foundation. I loosely characterize these elements as the who and the why. Examples of specific elements at this level include an opponent’ s goals, preferences, and perhaps even depth and breadth of resources.
- Operational elements come next and build on the foundation of strategic elements. I characterize the operational elements as the how and the what. Examples of elements at this level include the method of attack, the class of target, and the opponent’s technological capability.
- Tactical elements come last and build on the foundation of strategic and operational elements. I roughly characterize the tactical elements as the when and the where. It also includes the specific who, or the individuals who execute the event. Examples of elements at this level include the time and place of the event.
What does this simple model imply? Foremost, it suggests that a red team will probably not anticipate elements of a higher level correctly if it misreads elements of a lower level. Conversely, a red team that correctly identifies elements of a lower level is more likely to anticipate elements of a higher level. For example, a team that correctly interprets an opponent’s goals and preferences is more likely to anticipate the target and method of an attack. It is also worth noting that successful anticipation becomes more difficult at higher levels; in fact, it is arguable whether a red team or analyst can anticipate the tactical elements based on analysis alone.
A spin-off benefit of the model is that it provides a shorthand heuristic for describing levels of surprise. Using this model, a strategic surprise occurs when an analyst or decision maker fails to anticipate all levels before an event; an operational surprise occurs when an analyst or decision maker anticipates the who and the why but fails to anticipate the what, how, when, and where; and a tactical surprise occurs when the analyst or decision maker anticipates the who, why, what, and how but fails to anticipate the when and the where. As always, the analyst should not overlook the possibility of deception. Successful deception at the strategic and operational levels is almost certain to enable a strong surprise.
Finally, the value of anticipatory analysis must be measured in concert with the function of intelligence collection. Even though standalone analysis is not likely to anticipate the tactical elements of the model above, good analysis at the strategic level can point intelligence collection toward the operational and tactical elements of the model. In other words, an analyst may never be able to anticipate the when and where of an unlikely event, but he or she might be able to identify the what and how with enough confidence to tip collectors to the domain of the where and the when.
- See, for example, Tetlock’s extended discussion in Expert Political Judgment. [↩]