For many of the dangerous situations confronting the United States today, there is precious little precedent and little guidance on how to respond. Look no further than the possibility that Russian President Vladimir Putin might use a nuclear weapon in Ukraine or that China might invade Taiwan. Add the prospects of a seventh nuclear test by North Korea, Iran acquiring a nuclear capability and widespread unrest in China … well, you get the idea.
The potential for global surprises has rarely been greater. And surprise is the enemy of any nation’s foreign policy.
The challenge for U.S. policymakers is to prepare for all these traumas and game out how the U.S. would respond — not just in the moment, but in prolonged and escalating circumstances. For example, not just whether Russia would go nuclear in Ukraine, but what the U.S. would do and what next steps Russia might take. All this to avoid relying on improvisation and potentially chaotic responses, if and when the moment comes.
The good news? Such planning is happening now when it comes to the Taiwan and Russian situations. The bad news? Those are as complex and dangerous as any scenarios in recent memory.
How does the scenario-planning work? Who does this work? My experience in government taught me that there are many ways to prepare for such uncertainties and entire teams of people whose job descriptions might best be described as “preparing for nightmares.”
The concept dates to the early 19th century and a Prussian army officer who took actual games to his superiors and suggested that they be used to simulate conditions on the battlefield. Today, war gaming in the U.S. happens inside government agencies and private think tanks, and it takes many forms.
There is the geopolitical war game — a kind of chess match of diplomatic moves and countermoves — and then there is the more kinetic variety, with U.S. military officers simulating teams and playing through a series of military “moves” in which the battlefield is constantly changing.
For years, the U.S. has used war games to simulate a U.S. war with China over Taiwan. As Grid’s Joshua Keating has reported, in classified Air Force war games held since 2018, “blue” teams representing the U.S. have repeatedly lost to Chinese “red” teams. That’s partly by design — these games are designed to highlight vulnerabilities — but the simulations have also highlighted specific issues involving China’s geographic advantages and its rapid development of certain weapons (in particular anti-ship ballistic missiles, capable of precision strikes on U.S. ships at a range of more than 900 miles).
The 2021 Air Force game reportedly showed improvement for the U.S. side, though the Air Force game commander highlighted the fact that many of the necessary U.S. military assets were not yet in development or production. His conclusion: “If we change, we can win.” More broadly, these war games have shown the grave damage China could inflict against Taiwan in the early days of a conflict, but also the likelihood of a long and drawn-out war once the U.S. was involved — a war that could be devastating to the U.S. and China both.
Other war games take into consideration both military capabilities and political factors, sometimes using a team that mixes government players with outside experts. This was the case in a war game in March run by West Point’s Modern War Institute, which simulated a Russia-Ukraine war just weeks after the actual Russian invasion.
That game opened with the U.S. players overestimating Russian capabilities but quickly coming to the prescient view that over time Russia could not sustain the combat power necessary to take and hold a major city. The game also foresaw the eventual need for a Russian mobilization, resulting political tensions in Russia and a long-lasting stalemate on the battlefield.
What’s gained in these games? In the Ukraine case, those conclusions helped underline some of Ukraine’s underlying strengths and, more importantly, to expose weaknesses in Russia’s position.
More broadly, the games help the U.S. “team” take the measure of itself, expose resource and coordination problems among U.S. agencies and with allies, and test how a range of responses might work under the pressure of time and surprise.
“Red teaming,” which also originated with the 19th century Prussian military, is an invaluable variation on this — and one that I’ve seen work effectively. Call it war gaming with a twist.
A team of experts is asked to “become” the country or group whose actions you are trying to anticipate. And “experts” is the key word. The team must consist of people with two qualities: deep expertise on the adversary, and an ability to challenge conventional wisdom and avoid “mirror imaging” (the tendency to assume the adversary will behave as Americans would). The team members must be expert enough to enter the enemy’s social, cultural and ideological milieus and think as they would.
Red teaming differs from war gaming in that there is no opposing side; you don’t want this group reacting to others — you just want them to replicate and channel the thinking, logic and planning of your adversary. In colloquial terms, to get in their heads.
So in my Taiwan example, this team would consist of people schooled in Chinese and Taiwanese culture and history, and ideally with fluency in Mandarin. They might be given two kinds of tasks: playing Chinese policymakers to game out how Beijing would pursue its aims on Taiwan, or playing Chinese officials reacting to setbacks in their strategy. In the Putin-nukes scenario, the same idea — but with the expertise focused on Russia, nuclear weapons and Putin himself.
I know red teaming can work, based on the CIA’s use of the technique after the terrorist attacks of 9/11, when the agency felt acutely responsible for ensuring that another attack did not take place. We took our most adventurous and unconventional thinkers and formed a red team with a very specific task: We told them to “become” the terrorists — to imagine how and where they might plot their next attacks. Combined with raw intelligence we collected, the red team’s work guided many of the steps the administration took to harden vulnerable targets in the U.S. and abroad.
This proved important because, in my experience, the U.S. tendency even after 9/11 was to rely almost passively on intelligence to warn policymakers; in other words, wait for a CIA warning and then react. Our message was that we would do our best but — against an enemy that played by no rules — there would always be a chance something might be missed and someone would get through. Better to augment intelligence warnings with proactive protection of potential targets that we could identify.
It was hard to convince Americans — government officials and ordinary citizens — what an open target the U.S. was at that moment. In the immediate aftermath of 9/11, watch lists were not yet systematic and effective; people could board planes with knives; and in the entire aviation system, there were only 33 air marshals. The red team exercise helped expose these and other vulnerabilities.
It’s a simple, well-known concept, one that appears to have had its origins in an ancient Catholic Church practice of testing the arguments for and against conferring sainthood. In its modern, secular form, it’s another tool that helps government decision-makers narrow the potential for surprise or error.
Devil’s advocacy is most useful when leaders have arrived at a conclusion with confidence but must test that conclusion relentlessly, given the disastrous consequences should they turn out to be wrong.
Unlike war gaming or red teaming, the key here isn’t military or cultural expertise; rather than getting into the adversary’s mind, the idea for a devil’s advocacy team is to clinically evaluate and challenge the mainline judgment — to test it by arraying the data in ways that arrive at a different conclusion, or to detect some missing piece that is distorting the judgment.
This may sound like an academic exercise, but it can have profound impact. When the CIA concluded that Osama bin Laden was in Abbottabad, Pakistan, in 2011, it used this technique to test its judgment. It asked another agency to look at the case and provide a competing assessment to see if it differed from the mainline view — in other words, devil’s advocacy. The agency also conducted what’s known as “competing hypothesis analysis” — asking what other conclusions could have been consistent with the analysis. Could the Abbottabad resident have been someone other than bin Laden? A different terrorist? A man who resembled him? Finally, several other individuals with no experience in the matter were brought in to do an additional cold review of the data and conclusions.
You might say that there were three rounds of devil’s advocacy, all looking for differences or weak spots in the analysis.
The precise results of those exercises remain classified, but I can say this much: The multiple competing assessments offered the Obama administration a 360-degree view of the information and analysis, and ultimately gave the president and his top advisers confidence that the data had been thoroughly scrubbed and tested as they wrestled with the decision to launch the special forces raid on Abbottabad.
It was a different case, Iraq in 2003, that encouraged the increased and more systematic use of such techniques — given the flawed assessments of Iraq’s weapons of mass destruction and the assumption that the U.S. could overthrow Saddam Hussein without the trauma that followed. As for the Iraq War itself, it’s not clear that any prewar devil’s advocacy exercise would have made a difference given the momentum for going ahead that had built up in the Bush administration and Congress.
“What if” analysis
Here again, the words seem simple, but in practice the technique is important. The idea in a “what-if” analysis is to shift thinking from “How likely is it?” (say, that Putin might use a nuclear weapon in Ukraine) to “How could it actually come about?” You start by assuming that the hypothetical has already happened and work back from there; you assume, for example, that China has attacked Taiwan, Putin has used a nuclear weapon or Iran’s regime has collapsed, and so forth. And as you work your way back, you consider what must have occurred at each step. What would be the indicators or tripwires to watch for? What would we expect to see if the nightmare scenario were coming? And how would we know?
And then you gear your technical tools and best agents to watch for those things.
Sometimes a global crisis or hot zone will defy nearly all the other tools. An outcome may be too uncertain, complex or controversial to place confidence in any single prediction. Here’s where “scenarios analysis” comes in. This was a technique I relied on often when I was managing what are called National Intelligence Estimates (NIEs) from 1995 to 1997 and wrestling with questions about North Korea, Russia, Iran, and nuclear and missile proliferation.
This is not about predicting the future or gaming out moves, but about mapping the range of possible outcomes. One of the most intelligence-savvy policymakers I worked with, Gen. Brent Scowcroft, national security adviser for President George H.W. Bush, often said the real role of intelligence was to “narrow the range of uncertainty when difficult decisions have to be made.” Scenarios analysis can help you do that.
Here, North Korea may be the most relevant — and nightmarish — example. Policymakers have worried for years about the possibility that a nuclear-armed North Korea could collapse under the combined weight of economic and social problems, grinding poverty, and a brittle dictatorship — and the nightmares that might follow. But the potential ways in which a North Korean collapse might play out are almost limitless; put differently, you’d have to conduct hundreds of war games or red-team exercises to plan appropriate responses.
So rather than war gaming, you start with the forces and events that could determine the outcome. You typically generate at least three or four scenarios — “best-case,” “worst-case,” and one or two in between, along with the indicators that each of those is coming to pass. Ideally, you make a judgment about the relative probability of these outcomes. This gives the government some guidelines for planning against various contingencies. In the case of Ukraine, your best case might be that Ukraine succeeds in pushing Russian forces off all the territory it has taken, including Crimea. Your worst case could be that Russia regroups and roars back to occupy most of the country. And you focus your energies on those that are most likely.
When it works — and when it doesn’t
At the CIA in the spring and summer of 2001, we had strong evidence that al-Qaeda was planning a major attack on the U.S. This was the result of a huge spike in intelligence reporting more than a conscious application of the foregoing techniques — although I think it’s fair to say that we were in a nonstop “scenarios analysis” exercise, weighing different possibilities in terms of potential targets and methods. We know it was coming and that it was coming to the U.S. But we were unable to identify timing and specific targets.
In the aftermath of 9/11, however, the response benefited from an elaborate “what if” exercise that had been carried out in the year prior. In its final months, the Clinton administration had given the CIA a task in the form of a question, based on the possibility of a major al-Qaeda strike on U.S. soil: What if the CIA and other agencies were unconstrained by resources and given special authorities in the aftermath of such an attack? In those circumstances, what would we do to destroy al-Qaeda?
Our response was to develop what we called a “Blue Sky” plan, a term that reflected the unconstrained conditions the administration had posed. We delivered the plan in December 2000; it came off the shelf at Camp David, where President George W. Bush had assembled the national security team for the first full discussion of response strategies four days after the 9/11 attacks. Two days later, the president told us to put our plan into action. Within weeks, CIA teams on the ground in Afghanistan had prepared the way for U.S. Special Forces — and with the combined effort of CIA and the U.S. military, Kabul and the Taliban fell by November of 2001.
In the case of the Putin-launches-a-nuke scenario, I imagine a mix of all these techniques would be needed: war gaming for the chess match of moves and countermoves; red teaming, to be sure that decisions are made with a sophisticated understanding of the current Kremlin mindset (no small task); and what might be called the mother of all what-if exercises — using that calendar-in-reverse approach to do everything possible to ensure that it never comes to that.
None of these exercises guarantees perfection — or anything close. But using them — and using the results wisely — may narrow the chances of confusion or failure in national security emergencies. If we can answer those questions about Putin, or game out all the military and political steps that might play out should President Xi Jinping move with force against Taiwan, we have a better chance of success if and when the nightmares come.
Perhaps the clearest wisdom on this comes from then-Gen. Dwight Eisenhower. Having commanded the largest amphibious military operation in history, the D-Day invasion of Normandy, Eisenhower was acutely aware of the many things that could surprise a decision-maker. He had no illusions that he could design a perfect plan that would hold up under the pressures and chaos of battle, but he also knew that not planning would leave him even more exposed to surprise and disaster.
As he put it: “In preparing for battle I have always found that plans are useless … but planning is indispensable.”
Thanks to Lillian Barkley for copy editing this article.