There are rare books that improve the understanding of the world in some fundamental way, and simultaneously are fun to read. For me, Black Swan is one of those books. Treat these notes as a tribute.

Firstly, who should read this book and why? Well, a good candidate would be someone who is in a role of taking decisions that have probabilistic outcomes with huge upsides or downsides. Or anyone who is trying to make sense of unpredictable things happening around us, like wars, market crashes, bestsellers etc.

Before Taleb digs deeper into the concept of Black Swan, he sets the context in terms of two kind of environments, Mediocristan and Extremistan. Mediocristan is the natural world we have been living in. The mean is relevant in this world, and as the deviation of an expected event from mean increases, probability of that event becomes extremely rare. In language of mathematics, the deviations follow a Gaussian curve. To better understand, lets do a thought experiment. We take all adult citizens of a city and make them stand in a line according to height. Say, mean height comes to be 5.7’’ (yes, this is an Indian city). Now if we just consider around 20% deviation from this mean i.e. number of people who are at 6.7” and 4.7”, the count of them will be very less. And at a deviation of 2 feet (40%), we would be surprised to find a single person. Now if tallest person in the world, Sultan Kosan (8.3”) joins this line at the end, how much deviation it will cause in the mean of the city? negligible. This illustrates the laws of numbers in Mediocristan.

Now as the people are standing in the line according to their height, we make a new announcement, and make everyone stand according to their wealth. Say, wealth of the mean person comes out to be 30K USD. Now we can easily imagine large number of folks at 45K USD (50% deviation), at 120K USD (300%) and many folks even at 1000% deviation. The positive outliers would be seriously affecting the mean already, and if you bring in Bill Gates, average person will be a millionaire. This is Extremistan.

Lots of natural stuff around follows the laws of Mediocristan, like height, weight, life expectancy etc. On the other hand, lots of man developed systems follow laws of Extremistan, like financial markets, wealth, book sales, social media subscribers, population of cities etc. Important human motivations like wealth and status both follow the rules of Extremistan, making it more interesting of the two.

We understand the world through some models, and these models are validated against the past observed values. Everything happening in this world, is somehow modelled in an explanation or an equation, which allows us to predict it’s future behaviour. An example of simple model is an ice-cream seller believing people eat thrice more ice-creams in summer compared to winter. Accordingly, she prepares seasonal supplies. Another example of a simple model is, a stock buyer believing that good earning means better stock price. Based on this, she takes action on exchanges. Equations in physics are basically models that predict the behaviour of physical environments. At a personal level, these models are a great blessing, as they help us not to overthink, and fast track the decision making. Our reliance on models is so much that many big man-made systems are now governed based on these models, think markets, economies, corporations, consumer preferences etc. Trust in models makes everyone expects that these systems would be resilient to the future events. But there is a catch, as these models are approximations based on observed behaviour, and not complete reality, what happens when reality works in a manner, unexpected or unpredicted by a model? Very often, it crashes the seemingly resilient system. The crash of Sri Lankan economy is the failure of their economic system, 9-11 was the failure of US’s defence system, 2008 was the failure of financial system, and black-holes are failure of physics as we know it.

Now if you have to pick, which category of systems will likely to have more Black Swans, Extremistan or Mediocristan? Obvious choice would be Extremistan. This is because our models are often great at predicting the mean behaviour, surprises comes at tail values with high deviation from mean. In Mediocristan environments, we saw that tails don’t cause big changes to the aggregate values of a system, but in Extremistan, tails can obliterate the aggregate behavioural expectations.

To a child, many things happening in the world may seem unexpected compared to an adult. This is because adult has a bigger repository of models owing to experience or studied learning. Even among adults, an event would feel surprising to a particular person because her knowledge of models is poor. But if an unpredictable event happens which isn’t coherent with all the models present with humanity, that would be a truly shocking event, a Black Swan.

Are Black Swans inevitable? Well, Taleb thinks yes. Since there are complex systems among us with extremistan properties, the models will always be inadequate and susceptible to tail event risks. Once a black swan occurs, the models get updated to incorporate the new data. But that updation is often curve fitting, and not a perfect fundamental understanding of the system, leaving risk of a new tail event possible. Author advocates the philosophy of empiricism, which says all knowledge is limited by what is observed phenomenon through senses or experiments, hence will never be a perfect representative of reality.

But we don’t always want to escape Black swans, for example a black swan like Google made Sequoia rich when most venture firms lost money in dot-com burst. So as an individual, we would want maximum exposure to positive black swans, and minimum exposure to negative ones. This is possible by engaging in activities that increase or decrease their respective odds. Lets take an example of Venture Capital. It’s an amazing business, because by putting bets on numerous early stage companies, the downside is quantified i.e. total fund size, but upside remains uncapped with potential of taking pie in unforeseen and lucrative business models. Compare it with someone hypothetical who provides insurance on market crashes. She will make predictable and limited money on 99% days, but can be obliterated on an unexpected day. At a personal level, there can be enormous examples, like not wearing a seat belt will be more comfortable on 99.999% days. It will be cheaper to live in a neighbourhood with high crime rate, but would be highly costly on 0.0001% of occasions.  

To resist Black swans, Taleb cites example from nature, like how our bodies have pair of ears, eyes, kidneys etc even if it is not the most energy efficient evolution. It provides the system sufficient redundancy to keep it alive in case of losing a vital organ. So redundancy is an important principle, not just for biology, but for business, economy, any system. In investing world, warren buffet’s Berkshire Hathaway always maintains billions of dollars in cash. They accept underperformance in bull markets to ensure survival in any bear market (if only, Srilankan government understood this).  Insurance is an another common mechanism, which individuals and systems follow to take care of unforeseen negative events.

Author also suggests to practice healthy skepticism. If anyone proposes a theory or explanation that can predict future behaviour of a complex system, whether it’s markets, elections or success of book, never take it on face value irrespective of how plausible it sounds. Use it, but be sceptical.

With regards to positive black swans, exposing oneself to maximum probability of them is important. No VC knows which is next Google, but meeting folks in San Francisco has 100x more chances of finding one compared to Chandigarh. Hence Taleb makes a case for living in big cities, and attending parties & events, even if they appear like waste of time.

In the end, we all love models of the world, because false predictability feels better than to be in a state where there is no predictability. But occasionally, love gives a big heartbreak.