Quote:
This sort of analysis has been popular among segments of both the left and the right over the past few weeks. In a rush to dismiss the problem, people compare the death toll of the coronavirus to that of the common flu, or perhaps with automobile deaths, cancer, diabetes, etc. While this is appealing to the rational brain, it makes the fallacy of comparing different risk types. More specifically, fat tailed vs. thin tailed risks, which author Nassim Taleb has famously discussed at length in lectures and books. The mathematics behind it gets quite complicated, but here's the gist to understanding why it's simply wrong to compare the fat-tailed risks of the coronavirus and thin tailed ones like the common flu or cancer.
-To illustrate different worlds of probabilities, imagine you randomly took two people and were told their total height was 12 feet. Based on the probability distribution in this scenario, the most likely result would be that each were 6 ft tall. It would be extraordinarily unlikely that one was 4 feet and one was 8 ft. Now imagine you took two random people and were told their total net worth was $30 million. Based on the probability in this scenario, the most likely result is that one is worth $30M and the other $0, it would be extraordinarily unlikely that each were worth $15M. It's clearly important to distinguish between different probabilities or your assessment of a situation could be way off.
-Another example is if you gave 100 people money to gamble with for a day, you'd expect a certain result, a small-moderate loss, but not a catastrophic one. There might be a few who lost everything, but some would win, and overall there's almost no chance of losing everything. Compare this with having just one person gamble for 100 days. The chances of catastrophe and losing everything are far greater. The risks are not the same, even though they could appear to be superficially. The first is a thin tailed risk, the second a fat tailed one.
-Often you'll year something like, “You are statistically more likely to die drowning in your bathtub than from a terrorist attack.” This makes the same fallacy as this meme, in wrongly comparing risk types. There is virtually no chance that bathtub drownings will double this year, let alone go up by 10X, it's a mostly known, thin tailed risk. But there is a chance that terrorist deaths could double, or go up 10-100X. There's also no chance of bathtub drownings causing a global recession or a war. Thus, it's completely rational to worry about terrorism more than bathtub drownings, as the two are radically different in their risk types. The response should also be different, it would be unwise to shift anti-terrorism spending to bathtub safety programs (at least assuming the anti-terrorism spending decreased the risk).
-The same is true of potential pandemics (fat tails) vs common diseases (thin tails). There's virtually no chance that common flu deaths will double this year, let alone go up by 10 times. There is a risk that the coronavirus deaths will go up 10, 100, 1000 times current levels. There's essentially no risk of cancer causing a global recession or overwhelming our medical system, there is of a pandemic. Comparing the two different risk types is wrong, and even dangerous. One event is known and can be controlled for, the other is unknown and potentially catastrophic.
