Franklin Roosevelt and Winston Churchill in 1943 in Morocco during World War II. Photo: AFP

In his Fireside Chat of February 23, 1942, in the bleak days two months before the Japanese were turned back at the Battle of the Coral Sea and then at the Battle of Midway, US president Franklin Roosevelt turned to George Washington for inspiration:

“For eight years, General Washington and his Continental Army were faced continually with formidable odds and recurring defeats. Supplies and equipment were lacking…. Throughout the thirteen states there existed fifth columnists – and selfish men, jealous men, fearful men, who proclaimed that Washington’s cause was hopeless…. Washington’s conduct in those hard times has provided the model for all Americans ever since – a model of moral stamina. He held to his course, as it had been charted in the Declaration of Independence.”

Roosevelt looked to Washington as a model to which the nation could aspire. The US was not geared up for war, but Roosevelt would call upon industrialists, workers and engineers to transform the nation into a well-equipped war machine. The nation had lived through the political struggles of the 1930s, and these were far from settled. But now was not the time for them.

The president put it bluntly: “We can lose this war only if we slow up our effort or if we waste our ammunition sniping at each other.”

He was calling on people to be mature, and to understand that only the morally bankrupt exploit serious crises for political ambition. By appealing to the Declaration of Independence, he was appealing to the principles that formed the nation, and to its founders, just as does every great nation. Japan and Germany were an existential threat, and Roosevelt appealed to the core moral foundation to reject the fifth columnists – tough words from a tough man.

There is no cost-free pity, such as “I feel your pain,” in Roosevelt’s words. Pain and death were the order of the day. The US was to suffer more than 400,000 military and civilian deaths in the war (for the Russians, it was 24 million). As did every great leader, he did not sigh before the enemy; rather, he challenged his own people: “In time of crisis when the future is in the balance, we come to understand, with full recognition and devotion, what this Nation is, and what we owe to it.”

The new coronavirus is not the Axis Powers of World War II, nor is President Donald Trump the same man as president Roosevelt. But this is a serious crisis – not just the dead of today, but the economic and political well-being in the years ahead, as well the future dead resulting from today’s decisions.

The choices of how to proceed cannot be approached via a scientific theory, because no such theory exists. There are only competing untested models that at best provide some approximation to the actual physical situation – how good an approximation being unknown. This is the case even if there are sufficient data to make accurate estimates of model parameters – and the situation is worse with insufficient data.

The political, economic, and medical leadership is playing an adaptive game, constantly changing and reversing course in the fog of war. Yet the immature and the unknowledgeable snipe at this or that decision, as if there were some optimal deterministic path to chart.

One can always go back and point to different decisions that would have led to better results. The issue is how those decisions would have looked at the time they would have had to be made, not after a hypothetical world has morphed into an objective reality. No doubt the US would be better off had President Trump shut down travel between the US and China on January 1. How many of today’s critics would have cheered?

Certainly Roosevelt would have been wiser to have committed a massive first strike against Japan well ahead of December 7, 1941. It was clearly a military mistake not to have done so. But was such a decision obvious on January 1, 1941? Would he have been acclaimed on January 2 for having done so?

One can debate how the costs of a struggle should be assigned, how their importance should be weighted, and how decisions should be optimized relative to those costs. Should the medical people have the major say, or the economists? Each group will see the situation from its own perspective, and should have input.

The choice of costs involves value judgments, and the situation is far too complex to write down a mathematical expression. There will be unresolvable debate. We must be concerned with mitigating near-term suffering and not overwhelming the hospitals, but not at the cost of much greater total suffering across a longer time frame.

Judgments must be made in an environment dominated by data-driven models, not physical laws, which are unknown. At best, predictions are probabilistic, and these depend on crude modeling assumptions.

Building highly complex models generally fails because of lack of knowledge and the difficulty of obtaining large numbers of accurate statistical estimates. The difficulty grows astronomically if one attempts cross-disciplinary modeling, such as merging medical and economic models – and this is what is needed.

Because of model uncertainty, decision-makers will disagree among themselves. They will be critiqued daily by other knowledgeable scientists and statisticians, and that’s all to the good. Competent criticism is constructive.

The public is shown a simple graph with two death curves, one peaking high and early under the assumption of doing nothing, and the other peaking lower and late under the assumption of lockdown. The situation is much more complicated. Even when only considering deaths attributable to the virus, there will very likely be multiple peaks.

Moreover, the real situation is multi-dimensional, including many other factors. For instance, there is the issue of diminishing resources for the long term, should variation in the behavior of the disease create equivalent or greater challenges down the road.

Perhaps most important, modeling the population as a whole without partitioning it into subgroups based on age, locale, and risk factors can lead to poorly performing intervention strategies.

Do any in the US media or Congress have a basic knowledge of statistical modeling – theory, not simply the ability to type some numbers into a computer? Absent such knowledge, they present trivializations of the true situation and snipe at decisions made in the risky game of data-driven modeling.

The United States (as any country) can allow itself to be torn apart by the immature and unknowledgeable, or it can follow Franklin Roosevelt’s advice and “understand, with full recognition and devotion, what this Nation is, and what we owe to it.”

Edward R Dougherty

Edward Dougherty is distinguished professor of engineering at Texas A&M University.