web analytics

Take direction from people who understand exponential trends, wash your hands and, when it’s over, remember which branch of history we went down.

Taking action according to the COVID-19 or the Coronavirus is difficult because epidemics / pandemics behave totally differently from the sort of phenomena that we deal with on a daily basis. As such I recommend combining the expertise of the best of our health experts with those who are familiar with extreme events and the taking of decisions under uncertainty.

How to Think

Thinking about network-dependent phenomena like the Coronavirus outbreak is challenging because they operate with non-linear scaling; in the case of this novel disease, this is exacerbated because we are operating under conditions of uncertainty. This combination is challenging not least because it is counter-intuitive.

To illustrate non-linear scaling, the now-cliche rice on the chessboard problem is helpful. Imagine starting by putting one grain of rice in the square on the corner of a chessboard, putting double this number (two) in the next square, double this (four) in the next square and so on for all 64 squares on the board.

Most people drastically underestimate the number of grains one would put on the board in total, which is 2^64-1 or 18,446,744,073,709,551,615 grains.

Many of the most important systems that we encounter operate with this dynamic, known as compounding: bank interest, company growth, the spread of memes, internet traffic. Systems of this type are continually misunderstood by people when relying just on their intuitions.

People put far less money into the bank than they should (because they don’t understand how much it could grow) and, to quote Bill Gates, “Most people overestimate what they can do in one year and underestimate what they can do in ten years.” (because they have trouble intuitively picturing how those disappointing gains of one year in business can be compounded over years and years into impressive returns).

The behaviour of the Coronavirus pandemic is confusing because it’s growth is exponential.

This exponential behaviour comes from the fact that everyone infected has the potential to infect new people. At risk of oversimplifying, imagine that a single person has the virus and infects two others; those others can themselves infect two others: now we are back to our grains of rice on the chessboard. See the exponential trend below, in green:

It is not surprising, therefore, that authorities and corporations have been slow and at times blind-sided by the virus.

Essentially, my advice can be only that the virus will move faster than you expect: so act faster than it feels like you need to.

To Whom to Listen

I am most certainly not a virologist or physician, but I do have a good feel for the exponential and chaotic when prompted, and have identified some good thinkers to listen to on the topic of the Coronavirus.

Conversely, experts seems to be just as prone to being blind-sided by exponential trends.

For example, a major publication came out with an article in January 28, claiming that the Coronavirus would be small beer compared to the Flu. Reach out to me in person if you want a link: I want to avoid finger-pointing, especially as many such articles were written in good faith and based on interviews with health professionals.

See the graph below from Worldometer, which depicts how Coronavirus cases were just about to skiislope upon publication of the above article:

See the graph below from Johns Hopkins. It breaks down cases by region: to the extent that we can believe the data from China, cases appear to be stabilizing, while they are increasing exponentially in the rest of the world.

Conversely, on January 26th, statistician and financial markets trader Nassim Taleb and two colleagues published a short paper with a number of observations and recommendations, among them:

  1. In January it was too early to know the R₀ of the virus, or the number of new people to which one infected person passes the virus: if we acted according to the average figure, we may have been under-reacting.
  2. Back then we also did not understand how long people could be infected before they showed symptoms, meaning the same risk of under-reaction and that these errors would multiply each other.

The Coronavirus was known to have the potential to spread exponentially, but just how many people had it, how fast it would spread and how many people would die were uncertain. Given this uncertainty, Taleb and colleagues recommended extreme caution: they are familiar with non-linear outcomes, and uncertainty.

Other useful guides to the situation include:

Here are some excellent podcasts on the topic (I recommend long-form interviews over news-bulletins, if you want detailed information):

Very smart people can still be fooled by exponential trends, not least because most people’s day-to-day lives, including those of most experts, don’t feature that much by way of radically varying outcomes or compounding. You file the legal briefs, you send the tweets, you fill in the TPS reports, you transport the passengers — no matter how hard you work, you’re unlikely to come near to changing your performance by an order of magnitude.

Taleb, however, knows that traders can blow up and lose more in a day than they made their firm during their whole career; Naval knows that most startups fail but a few make mega-mega-bucks. We need to learn more from people like this, about this genre of thinking: understanding the rare, the non-linear and, in this case, the highly risky.

How to React when It’s Over

To lean very heavily on Taleb, again, in his book The Black Swan, he describes an excellent thought experiment: imagine that, some time in 2001, the USA passed a law that all airlines should fit their aeroplanes with re-enforced doors, separating the cockpit and cabin. Assuming the airlines were expected to take timely action, this would have foiled 9/11.

Would we regard the legislators who drafted the law as lifesaving leaders of prescience and intelligence? Unlikely. The terrorists would probably have given up and tried something else. This is to say that sensible preventative measures, if they work, can diminish the sense of their own importance.

To the extent that the worldwide efforts to fight the Coronavirus are successful, people who opposed caution and swift action will make the claim that it was a waste of time. Absent 9/11-style attacks, so might airlines in our thought experiment.

The branching universes visualization is useful here: we should think about the Coronavirus and other similar situations in terms of a tree with branching timelines based on different choices. What would have happened if we did nothing? If we did much more, sooner? This and other options.

The 9/11 thought experiment makes this easy because it’s a binary: was the door terrorist-proof? In the case of the Coronavirus, imagine the following:

  • A branching tree of possibilities each of which corresponding to different values for the transmission rate, the mortality rate, reinfection, the length of time during which patients are asymptomatic, etc.
  • Add to the above, branches for each action taken/recommendation issued by experts and non-experts, and when: quarantine, isolation, testing, etc.

When this is all over, many will succumb to the temptation to call people saviours, to assign blame, according to whether they made the “right” decision (with hindsight) — as though we weren’t making decisions in a condition of high risk and uncertainty. It’s possible, of course, to do the wrong thing but for good reasons given the available information, and vice-versa.

Instead we should orient ourselves toward the situation with the best possible understanding of the risks involved and of the risks associated with action.

Leave a Reply