What is a black swan event? Are black swan events truly unpredictable or just highly improbable?
A black swan event is an extremely rare and unpredictable event or occurrence that has major negative consequences. The concept is based on the idea that someone who has only ever seen white swans would naturally assume that all swans are white and would therefore be completely unable to predict the existence of a black swan.
Keep reading to learn about the concept of a black swan event.
What Is a Black Swan Event?
The notion of a black swan event has been developed by a mathematical statistician Naseem Taleb to denote major, history-shattering events that could not have possibly been predicted. But if these events cannot be predicted, is forecasting a fool’s errand then? For that to be true, we need to accept both of Taleb’s conclusions: that black swan events are literally impossible to predict, and that only black swan events change the course of history.
Is “A Black Swan Event” Truly Unpredictable?
If the term “black swan event” only describes truly unpredictable events, then there have been very few actual black swans in history. The most commonly cited black swan example is the 9/11 terrorist attacks. In that example, catastrophe literally fell from the sky without warning. But even 9/11 was not completely impossible to predict—similar attacks had been thwarted in the past, and the intelligence community was actively examining the threat. In fact, in 1998, the Federal Aviation Administration even analyzed a possible scenario in which cargo planes were hijacked and flown into the World Trade Center.
(In practice, that evidence means that while it would have been extremely difficult to predict the exact date and time of the 9/11 attacks, it was entirely possible to predict a terrorist attack in which a plane was turned into a flying bomb, and several people actually did make this prediction before the event.)
But what if we take “a black swan event” to mean, in Taleb’s words, “a highly improbable consequential event” rather than a event that can’t possibly be predicted? In that case, Taleb’s logic becomes easier to swallow.
But black swans, even defined more loosely, are still incredibly rare by definition, and it could take hundreds or even thousands of years to generate the amount of data that would allow us to calibrate how accurately we can predict them. In that sense, Taleb is correct—trying to predict a black swan event is pointless.
Are “Black Swans” the Only Events That Matter?
The second part of Taleb’s logic is that only black swan events change the course of history. This claim is much easier to dispute—the gradual development of technology and slow growth of the global economy have had enormous consequences. The gradual increase in life expectancy is a result of those changes as well as advances in medicine, hygiene, infrastructure, and so on. Living longer than previous generations is an undeniably important development, but it has no single black swan cause.
Another way to think about the value of black swans is through the lens of investing. Taleb’s beliefs about unpredictability are part of his massive success as an investor. In volatile fields like technology where most startups fail, it’s almost impossible to predict which lucky few will survive and go on to become the next Google or Amazon. For Taleb, wild success is a black swan event (which is unpredictable by definition), so he casts a wide net of investments rather than sinking time into analyzing any particular option.
This is “black swan investing.” It’s similar to buying lottery tickets—most will not be winners, but the more you buy, the more chance you have of holding the one ticket that wins big. Black swan investing is high risk, high reward.
But black swan investing is not the only way to make a profit, and a high-risk strategy is not for everyone. Many people have made similar fortunes by doing their research and making informed predictions about a company’s likelihood of success. The profits may be smaller than in black swan investing, but the losses are smaller too, and over time, both approaches are equally lucrative. In investment as well as in history, black swans are important, but they’re not the only thing that matters.
The Consequences Are Predictable, Even if the Event Isn’t
It’s also important to consider where we draw the boundaries of a black swan event itself. For example, we don’t talk about the 9/11 terror attacks as a black swan event just because they were more or less unpredictable—unpredictable things happen all the time without earth-shattering ramifications. But 9/11 marks a turning point in world history because it launched a decade of armed conflict with casualties in the hundreds of thousands, not to mention a total overhaul of air travel security. In other words, what makes black swans so important is not just the event itself but the consequences.
This presents a challenge to Taleb’s logic. While the details of the 9/11 attacks themselves may have been unpredictable, the consequences were not. Superforecasters have made accurate forecasts on questions like the likelihood that one country will invade another (like the U.S. invaded Afghanistan) or that a terror suspect will flee to another country to escape that invasion. If black swans are important in part because of their consequences, and superforecasters can predict some of those consequences, then forecasting itself is a useful enterprise.
How Do We Prepare for the Unpredictable Black Swan Event?
Taleb, Kahneman, and Tetlock all agree that trying to predict events more than a few years in the future is pointless because too much can change in that amount of time. But not anticipating future events is often not an option.
For example, although it’s impossible to predict the exact timing of future earthquakes, building codes in earthquake-prone regions still mandate that all new construction be engineered to be as earthquake-resistant as possible. It’s possible that advanced technology would never be needed. But the massive devastation a large quake can cause means the risk is worth the investment.
Similarly, for many years, U.S. military protocol mandated that the armed forces always maintain the resources to fight two simultaneous wars. It’s impossible to know for sure whether that would ever be necessary—but just like with earthquake-resistant buildings, the risk is too high not to prepare for the possibility.
All of this preparation depends on context. Fighting two simultaneous wars would be a terrible turn of events, but it’s not the worst possible scenario. So why prepare for two wars instead of three or four? And why do building codes in Florida not put as much emphasis on earthquake technology as building codes in Tokyo? In both cases, the probability of disaster must be high enough to justify the cost of preparation. The likelihood of a three-war scenario or a megaquake in Florida is so low that it does not justify the enormous cost of preparing for those scenarios.
Non-Normal Probability Distributions
In order to know whether preparing for the worst-case scenario is worth the investment, we need to know how probable that scenario is compared to other possible outcomes. To do that, we need to know how those outcomes are distributed. The bell curve in Chapter 3 represents the normal distribution of a given trait within a population. For example, height is normally distributed, with most people clustered around the average and an increasingly small number of people at the extremes.
But some variables don’t fit the bell curve. For example, household wealth in the U.S. is not normally distributed. If it were, there would be virtually no billionaires, since one billion is at the very extreme tip of the curve, as far from the median as it is possible to be. In reality, the likelihood of a random household having a net worth in the billions is one in every seven hundred thousand. This is because household wealth follows a fat-tailed distribution, where the average is skewed to one side of the median, allowing the opposite tail to remain “fat.”
On the graph below, the grey region represents the normal distribution and the blue line is a fat-tailed distribution. Notice how the blue line stays far above the X-axis on the right side. This means that a higher than normal number of people will fall on the high extreme of this variable.
What does this have to do with the value of forecasting? Probability distributions are important because they dictate our ability to make accurate predictions. For example, if you assume net worth is normally distributed, your prediction about the number of billionaires in the U.S. will be way off. They also give us a way to think about counterfactuals—“what if” scenarios and alternate histories that didn’t happen but were once equally possible.
If we assume counterfactuals are normally distributed, then forecasting makes sense since extreme deviations from the mean (the prototypical “black swan”) would be incredibly rare. However, Nassim Taleb argues that counterfactuals actually have a fat-tailed distribution. If that’s true, the odds of an extreme event like a world war go from one in trillions to one in hundreds of thousands—still not likely, but far from impossible.
Does Forecasting Matter?
We look back on counterfactuals as alternate histories, but to someone living at that moment in time, they were alternate futures. This means that our own possible futures are also wide open, and extreme events might be far more likely than we think. But even if Taleb is right on all accounts—that the distribution of possible futures is fat-tailed, making history-shaping black swan events far from impossible—it still does not negate the value of accurate forecasting.
At the very least, accurately predicting the consequences of unforeseen events can be a valuable tool for coping with the aftermath of disaster. And while black swan events do create profound change, they’re not the only way those changes come about. Talented forecasters who can give reliable predictions even slightly better than chance will always be a valuable asset.
———End of Preview———
Like what you just read? Read the rest of the world's best book summary and analysis of Philip E. Tetlock's "Superforecasting" at Shortform .
Here's what you'll find in our full Superforecasting summary :
- How to make predictions with greater accuracy
- The 7 traits of superforecasters
- How Black Swan events can challenge even the best forecasters