Review: Fooled by Randomness by Nassim Nicholas Taleb

Interesting content, horrible writing. Early in the book, Taleb mentions that he ignored most of the feedback from his editors; the result is a book that’s poorly organized, has awkward sentence structure, seems to wander randomly (haha) from topic to topic, and uses a voice that is horribly smug. Taleb insults every “expert” everywhere—especially investors, MBAs, editors, and journalists—and is all too happy to tell you how everyone is an idiot and wrong about just about everything. It’s off-putting and made me want to put the book down many times.

The only things that kept me going were (a) that I had read many rave reviews about this book and (b) I was able, after much skimming to get past the obnoxious attitude, to find a few interesting insights here and there. Some of the main ones:

• The key argument of the book is that far more of life (but not all!) is attributable to random chance than we normally assume. It’s human nature to come up with elaborate stories of causality and hero narratives, but this is largely due to a variety of cognitive biases, especially hindsight bias and survivorship bias.

• Given a large stream of data (e.g., stock data), the shorter the time scale you use when looking at the data, the more noise you’ll find. For example, the more often you check your financial portfolio, the more likely that the up and down moves you observe are purely random, rather than any meaningful pattern. Over longer time periods, the fluctuations tend to cancel out and the real patterns emerge (law of large numbers).

• The larger the pool of candidates, the more likely it is that the success of some of those candidates is due to pure randomness. For example, if you have 10 people trading stocks, and one does significantly better than the average (especially over a longer time period, as from the previous point), then it’s likely due to that person’s skill. But if you have 10 million people trading stocks, and one does significantly better than the average, then that is much more likely due to dumb luck.

• In fact, you can model this: start with 10 million people and have 5 million randomly pick one stock position and the other 5 million pick the opposite position. After a few weeks, one of these groups will have made money and the other will have lost money. Take the group that made money, and now have 2.5 million of them randomly pick another stock position and the other 2.5 million pick the opposite position. Keep repeating this experiment, each time taking the group that made money and sub-dividing them, and after 10 repetitions, you’ll have roughly 10,000 people who made correct stock picks 10 times in a row—and made loads of money—through pure randomness.

• Based on the two items above, it’s worth noting that the most successful traders in any (relatively short) time period are often the worst traders! They are the ones with no understanding of what they are doing, who just happen to blunder into success through pure random chance. However, over multiple cycles, their luck will run out, and they will also be the ones to “blow up” in the most spectacular fashion.

• We are used to systems where more information means more certainty about your knowledge: e.g., if you’re picking marbles out of a jar, each new marble you get increases your confidence in knowing the distribution of colors in that jar. However, there are many systems in life where this isn’t the case: e.g., perhaps the marbles are not evenly distributed (e.g., all the red marbles are at the top, all the blue at the bottom) or, worse yet, perhaps a mischievous actor is sneaking marbles into the jar when you look away (e.g., after you took out 100 red, they dump in 100 blue).

• Probability is not same as expected value, though most people confuse the two. Having a low probability of losing money is good, but if that loss ends up being catastrophic, then the expected value is still going to be very low, and may be worth avoiding.

• Real life is like the infinite monkeys at infinite type writers idea, except we only see the monkeys that (randomly) produced Shakespeare and not the countless more who failed to do so. E.g., We typically only see the successful professional athletes on TV and not the vastly larger number of athletes that never made it. Since we typically only see the winners, we vastly miscalculate the odds of success and misunderstand what it really takes to be successful (e.g., we may notice all professional athletes have coaches and assume that’s what it takes to be successful, not realizing that all the athletes who never made it to the pros may have also had coaches).

• The cycle of life: you start off poor, then you get rich, then you move to a more affluent neighborhood, and now you’re poor again.

• True randomness does not look random to us. Consider dart board with 16 squares. If you threw 16 darts completely at random, the odds of getting 1 dart in each square is very, very low. Instead, it’s much more likely that some squares will have several darts while others have no darts. This sort of “clustering” is totally normal in random data, whereas uniform distributions are exceptionally rare.

• Much of life is the result of “path dependance:” that is, where we get a result not due to its merits or some logical decision, but merely as the next step from some previous result. For example, we use the QWERTY keyboard layout on almost all computers not because it’s the most efficient, but because it’s one of the least efficient! This is due to path dependence, where old mechanical typewriters used to jam if you typed to fast, and the QWERTY layout was developed to reduce typing speed. It stuck, and due to the way history developed, it’s nearly impossible to change it now, even though better layouts exist.

• Random effects are like compound interest: a small random win early on lead can lead to a disproportionately huge advantage later on. This is largely due to path dependence. For example, consider two children, Bob and Steve, who both take up hockey at age 8. Let’s say before their first game, Bob happened to have a great night of sleep, while Steve slept poorly. As a result, Bob plays great, whereas Steve has a lackluster performance. Going into the second game, Bob is now feeling confident, and plays well again, whereas Steve feels outmatched, and plays poorly again. After a year or two of this self-reinforcing loop, Bob may end up being picked for the “A” team at school, whereas Steve get stuck on the “B” team. The A team is seen as the more hopeful candidates, so they get more practice time, better coaches, and stronger opponents than the B team, so gap in skill between Bob and Steve will increase further. Over a number of years, this will compound further and further to where Bob may end up going to the pros, whereas Steve gets cut from the team in high school—all due to a totally random occurrence very early on.

Having read all this, and found some of it interesting, it’s not clear what we’re supposed to do with it. Yes, there is a lot of randomness in the world, and yes, we’re all largely blind to it based on how the human brain evolved. But so what? Calling everyone an idiot and throwing away all the tools and techniques we use today doesn’t seem like a good solution (I’m reminded of the quote, “All models are wrong; some models are useful”). I suppose the main contribution of this book is increase awareness of randomness, which is definitely a worthwhile activity, but I would’ve liked to see more concrete advice from Taleb on how to better deal with it (beyond the 2-3 “tricks” he mentions in te final chapter).

As always, I’ve saved a few of my favorite quotes from the book:

“Reality is far more vicious than Russian roulette. First, it delivers the fatal bullet rather infrequently, like a revolver that would have hundreds, even thousands of chambers instead of six. After a few dozen tries, one forgets about the existence of a bullet, under a numbing false sense of security. Second, unlike a well-defined precise game like Russian roulette, where the risks are visible to anyone capable of multiplying and dividing by six, one does not observe the barrel of reality. One is capable of unwittingly playing Russian roulette - and calling it by some alternative “low risk” game.”

“My lesson from Soros is to start every meeting at my boutique by convincing everyone that we are a bunch of idiots who know nothing and are mistake-prone, but happen to be endowed with the rare privilege of knowing it.”

“There is a simple test to define path dependence of beliefs (economists have a manifestation of it called the endowment effect). Say you own a painting you bought for $20,000, and owing to rosy conditions in the art market, it is now worth $40,000. If you owned no painting, would you still acquire it at the current price? If you would not, then you are said to be married to your position. There is no rational reason to keep a painting you would not buy at its current market rate—only an emotional investment. Many people get married to their ideas all the way to the grave. Beliefs are said to be path dependent if the sequence of ideas is such that the first one dominates.”

Rating: 3 stars