Dealing with Uncertainty in Strategic Decision-making

Abstract

ABSTRACT: Strategic uncertainty is the disparity between what one knows, and what one needs to know in order to make a responsible decision; it permeates defense decision-making. Because of strategic uncertainty, planners must maximize the robustness against surprise in striving to achieve critical goals. This article describes the decision methodology known as robust-satisfying and the integration of this method with other military decision-making processes. ********** Flipping a fair coin has equal chance of getting heads or tails. Rolling a balanced dice has equal probabilities for each of six A known outcomes. But if we take this into the realm of strategic decision-making and consider the 2002 assessment of Iraqi capability with Weapons of Mass Destruction (WMD), how many outcomes should we ponder: what are they and what are their likelihoods? One might say the answer is binary: Either they do or they do not have WMD. Or, perhaps we should consider multiple possibilities: They have small (or large) quantities, they are (or are not) developing more, and they intend to use it (or not). It is as though we are rolling dice without knowing how many faces each die has, and whether or not each is balanced for equal probabilities of all outcomes. This is essentially the problem every strategist faces, and the one this article proposes to address. We often are justified in thinking probabilistically and in saying something is very likely. For example, Stalin's military advisers in 1941 claimed a German invasion of the Soviet Union was very likely. The advisers had reconnaissance evidence, captured documents, and more. (1) Most analysts (though not Stalin) readily acknowledged the complementary assertion--Germany is not about to invade Russia--was very unlikely. In binary logic, an assertion is either true or false. If we know an assertion is true, then we know the negation of that assertion is false. There is an in binary logic. The excluded middle rules out the possibility an assertion is both true and false. Probabilistic thinking is an extension--to the domain of uncertainty--of the binary thinking of pure logic: If we know an assertion is highly probable, then we know the negation of that assertion is highly unlikely. An assertion and its negation cannot both be highly when using probabilistic reasoning. In strategic affairs, we often do not know enough about the situation to exclude the middle as we routinely do in binary logic and in probabilistic thinking. The British during World War II could have viewed the assertion that Germany was trying to build an atomic bomb as likely (indeed they were). Otto Hahn, who was a war-time professor in Berlin, had visited Enrico Fermi during the latter's experiments with uranium in the 1930s, and Hahn won the 1944 Nobel Prize in Chemistry (awarded in 1945) for his discovery of fission of heavy nuclei. (2) But one could argue the Nazis abjured Jewish physics, such as relativity and quantum theory, and therefore it is unlikely Germany would try to exploit this physics in order to build an atom bomb. Indeed, the Nazis never pursued nuclear weapons as enthusiastically as the Allies. If one needs to say an assertion is both quite and quite unlikely, one must abandon the binary structure of probability. This need arises quite often in strategic affairs. One reason is conflicting intelligence reports are common, as the Prussian military thinker Carl von Clausewitz emphasized. (3) Another reason is we often are unaware of, or do not understand, new doctrinal or technological possibilities. For instance, the possibility and implications of massive infantry use of hand-held Sagger anti-tank ordnance surprised the Israelis in the Yom Kippur War, despite their experiences with similar missiles both as users and as targets. (4) Furthermore, prediction is always difficult, especially in war. …