The Black Swan is a non-fiction book from the series "Incerto" by polymath Nassim Taleb. It’s subject is a combination between philosophy, finance, probability and a bit of psychology. Taleb centres the book around a construct he calls Black Swan events (capitalized). Named after a famous philosophical thought experiment “The problem of induction”, Taleb defines a Black Swan event as one that is highly improbable, has drastic consequences and retrospectively appears explicable by narrative. He details throughout the book how these events dominate the course of history and how a lack of understanding of Black Swans causes us to misunderstand Black Swans and by extension, how the world works. This author is highly controversial and generates a lot of criticism, mostly for his style which is informal and dismissive of those he disagrees with. While I agree with that characterization, I urge the reader to look past this to access the ideas which are original and insightful. The thesis of this book has many implications and I recommend to anyone seeking a probabilistic understanding the world should read this book and come to their own conclusions regarding its arguments.
In Part one he defines some concepts that are pivotal throughout his series. I will probably refer to them as well as many of them are really useful.
The first one I will mention is he categorizes variables into either "Mediocristan" or "Extremistan".
Mediocristan: A variable that is said to be from mediocristan is bounded in terms of the possible outcomes and generally follows Gaussian distribution. Individual events cannot change the overall behaviour of a mediocristan variable e.g. no matter how many calories you eat today, you will not significantly change the overall caloric consumption of your year.
Extrimistan: A variable is said to be from extremistan if its outcomes are unbounded and follows a Mandelbrotian distribution. Single events can drastically change the behaviour of the collective e.g. you cannot get an accurate picture of the Mexican economy without including Carlos Slim (who accounts for 40% of the Mexican Stock Exchange).
One of the main arguments of the book is that our minds are not geared towards extremistan variables which causes all sorts of problems with how we make predictions and assess evidence. For example, I may refer to the Turkey Problem from time-to-time, partially because its name is funny...
The Turkey Problem: consider a turkey who lives on a farm. He is fed by the farmer everyday for 1000 days and has come to the conclusion that “farmers are good for turkeys”. However, day 1001 is actually thanksgiving and the poor turkey will be caught by surprise. Not only were the 99.9% of instances misleading, they led the turkey to believe the exact opposite of the truth. This is a great demonstration of the Black Swan theory: 99.9% of the time farmers are good for turkeys so clearly a slaughter is “improbable” from the turkey’s perspective, the slaughter has a drastic impact on the turkey and retrospectively it would be obvious that the farmer was raising the turkey for slaughter on thanksgiving.
Here is an idea that isn't explicitly named that I will call the Tsunami Fallacy.
Tsunami Fallacy: In planning, people tend to look to past events for a scope of the problem e.g. when building earthquake/tsunami protective measures, they consider what is the largest tsunami ever recorded. The problem is that at the time of that largest event happening, it by definition, was unprecedented in magnitude. Surely enough Japan was eventually hit by an earthquake with larger magnitude and because their precautions were insufficient, it did considerable damage. People (including “scientists”) mistake absence of evidence for evidence of absence. I personally experience this all the time in research journals where people say “no evidence of X was found” and use that to mean or claim that they do not believe X exists.
This part of the book is about how "We Just Can't Predict". Taleb applies many of Daniel Kahneman's ideas (See earlier post "Thinking Fast and Slow") to explain how we have a false sense of understanding past events. He talks about a study where women were shown a bunch of identical stockings and yet were able to provide reasoning why they liked one better than the others. There was another study that showed how in people whose left and right brains are disconnected, they were able to get one side of the brain to make up stories to explain the other side's actions. The many examples of psychology like this show us that we will make up reasons so that we feel as though we understand.
He also talks about how we have no understanding of probability and our confidence levels. Participants in a study were asked to give a 98% confidence interval (a range of answers that will be correct 98% of the time) yet their error rate approached 45% even after a lesson on confidence intervals. The conclusion was that people can't accurately correspond confidence to probability.
Taleb also challenges the "gen-pop" understanding of Chaos Theory by examining the classic physics problem of a table of billiard balls and trying to calculate their eventual position using the conservation of momentum. Supposedly, we can scarcely accurately predict even one collision, but to do the second collision we would need to account for each player’s gravitational pull and to do all Nth ball collisions you’d need to account for every electron in the universe. While I haven't had a chance to review his calculations on this, I can at least understand the principle that very small uncertainties accumulate to large uncertainties over time.
Together, the false sense of understanding compounded by how large uncertainties grow over time, Taleb argues that we cannot predict things as well as we think. Even master predictors such as Warren Buffet, he says, are statistical phenomena where if you have enough people, some will get lucky a lot of times in a row. Paying too much attention to these individuals is therefore a form of survivorship bias.
He discusses a solution to these problems called "the barbell strategy" which will be addressed in a later post on the Black Swan's sequel "Antifragile".
In this section he discusses how the probabilities of the so-called "80/20 rule" actually have really profound implications on many facets of life. The proper name for the 80/20 rule is the Pareto distribution and simply put, 80% of the effects are accomplished by 20% of the cause. This applies to many variables that have to do with productivity and success. For example, according to this distribution, 80% of money is owned by 20% of people. If we extend this rule to the extreme case we find that the top 1% would account for 50% of the effects.
He also defines a “grey swan”:
A grey swan is an event that is possible and known, potentially extremely significant, but is considered not very likely to happen.
They can be positive or negative and significantly alter the way the world operates, which is why we are urged to take them seriously.
Examples of grey swans include climate change, population growth, and rising debt.
As always, after I read a book with lots of rich information, I ask myself how I will convert this into tangible improvement in my life. With The Black Swan, the ideas are absolutely relevant to the real world but the value is not in physical things you can do but in ideas you should use to filter incoming information.
The most obvious application would be to modify ones risk taking strategy. If you are investing your money, time, or any kind of resource then you want to do it wisely. Someone who is aware of Black Swans can strategically choose whether or not to put their chips on things from extremistan. They might prefer for example to invest their efforts into software development, which is from the capitol of extremistan (just look at Facebook vs the startup graveyard). But the same people might prefer to put their money in something extremely mediocristan--Taleb gives the example of investing in the US treasuries. [disclaimer: I am not a financial expert and this should not be taken as financial advice] Together this could ensure a slow and steady growth in your investments while the software development provides the possibility of being the next Mark Zuckerburg.
Another really common application of this book's ideas would be to be more careful while consuming the news. An entry into a news outlet typically has two, sometimes three components. First is a descriptive part where they explain what has happened. Second is an analysis which places today's news into a (usually much larger) narrative of an ongoing story which explains it all. The final component, if the author should be so brazen, is to tell the reader what they ought to think about the subject and its narrative. A proponent of the ideas from "The Black Swan" would like be wary of the journalistic (and human) tendency to craft an overly simplistic narrative that allows us to make sense of complex systems--which are almost universal in the contents of the news. When consuming the news I do my very best to only listen to the descriptive phase which is really the majority of the value. (If you're interested in the psychology of consuming news, I will be commenting more extensively during a future review of "Enlightenment Now")
Someone who understand power-law distributions and extremistan variables will appropriately analyze such distributions. For example, they will not be surprised by high levels of inequalities, a common example I have heard is that the 1% own 50% of America's wealth--but Taleb would cite that the 80/20 rule could just as easily be called the 50/1 rule.
I could go on and on about the implications of the Black Swan hypothesis, but thankfully I don't have to. Taleb did that himself and called it "Antifragile" which is the next book in the Incerto series. I will likely cover that book next since the ideas are so deeply intertwined with the ones here. Some exciting ideas I haven't mentioned yet are the fractal property, asymmetry of Black Swans and of course Antifragility.