Chapter 5: Survival Of The Least Fit—Can Evolution Be Fooled By Randomness?
The author tells the story of two traders who, over a few years made an enormous amount of money, only to be wiped out in a crash in a few days.
These people weren’t good investors. They were simply at the right place at the right time, trading the right instruments.
They were fools of randomness.
Here are their traits:
- An overestimation of the accuracy of their beliefs in some measure, either economic or statistical.
- A tendency to get married to positions
- The tendency to change their story: going from short-term trader to “long-term” investor
- No precise game plan ahead of time as to what to do in the event of losses: they never even imagined they would lose.
- Absence of critical thinking expressed in absence of revision of their stance with “stop losses”: they refused to use stop losses.
- Denial
The best traders are often seen as the worst ones in the short and medium-term, but are revealed to be the best ones in the long term.
The worst traders are the best in the short and medium terms, but the worst in the long term. They are the ones making huge gains due to taking huge risks. Until time catches them, and they end up going bust.
This is counter-intuitive because of Darwin. According to Darwinian theory, companies improve because they compete. As time passes by, the bad companies go bankrupt. Only the good ones survive.
But this is false, for two reasons. First, Darwinian selection is based on reproduction, not survival (species improve when they give birth to a better version than their current ones, not when they grow old).
Second, randomness can negatively influence genes.
Steven Jay Gould demonstrated the existence of “genetic noise”. That is, that some species “genetically regressed” due to randomness. Inferior genetic traits, despite being inferior, still managed to survive.
However, this was only in the medium-term. Over a long period of time, species improve (or disappear). Darwin’s theory was about the long-term.
-> we don’t live in a world where things are continually improving.
In fact, things don’t even move continuously.
If we stretch time to infinity, every species will end up disappearing. Darwinian evolution does not mean that all species are at the pinnacle of their form, all of the time.
Similarly, a bad trader that is winning seems like a great reproductive partner for the medium and short term. In reality, he isn’t for either of them.
Chapter 6: Skewness And Asymmetry
Let’s talk about asymmetry with gambling as an example. Asymmetric odds mean that the probability for each event is not 50-50, but say, 40-60.
Asymmetric outcomes mean that the payoffs are not equal.
Imagine that you play a game where you have 999/1000 chances to win $1 (the probability), and 1/1000 chance to lose $10 000.
| 999/1000 = 0.999 0.999*$1 = $0.999 -> this is your expectation | 1/1000=0.001 0.001*-$10 000= -$10 -> this is your expectation |
-$10 + $0.999 = -$9.001
Your expectation at this game is that you lose -$9.001.
Probability ≠ expectation.
Bull and Bear Zoology
Bullish and bearish mean little when we prefer the expectation to the probability.
If there’s a 70% chance that the market goes up by 1%, but 30% that it goes down by 10%, the expectation is:
- (0.7*1%) + (0.30*-10%) = 0.7 – 3 = -2.3%.
The expectation is that the market goes down by 2.3%.
To quote the author:
It is not how likely an event is to happen that matters, it is how much is made when it happens (…). How frequent the profit is irrelevant; it is the magnitude of the outcome that counts.
As a result, you can be bullish on the market but still have negative total expectations.
The author explains then he never sought to make money regularly on the stock market. Quite the opposite.
He made money on “rare events”. These events were unlikely to be predicted and unlikely to happen twice. He did so because these events were widely undervalued – which means he could make a lot of money at one time.
This is also why the author does not predict the future based on the past. Because you have in history, events “that never happened before” but that will happen in the future.
We don’t know about these events. Hence, using the past (where they did not exist) to try to predict the future is useless.
All things that never happened eventually do happen.
The broader the period of history you look at, the better it is. The problem is that people look too much at recent history.
The Rare-Event Fallacy
The author defines a rare event where the quote “beware of calm waters” can be applied.
It is, for example, the polite and discreet neighbor that ends up being a serial killer.
The author associates rare events with any misunderstanding of the risks derived from a narrow interpretation of past time series.
A rare event exists precisely because it is unexpected. If it was expected, it would not have happened.
Statistics is based on the idea that the more info we have, the more we can confidently predict a result. It’s true, but only when distribution is symmetric.
If you have an urn with 10 black balls, but ignore whether there are red balls or not, you will find out whether there are red balls much faster than if there isn’t any (you need to draw the 10 balls to be sure there aren’t any red balls, while drawing one red ball would confirm that there is at least one).
Asymmetry in knowledge is extremely important.
Imagine the stock market is an urn with a set number of black and red balls. Now, imagine that someone replaces the red balls with black balls without you knowing. Suddenly, the fundamentals of the urn change, and so will the expectations.
In that case, looking at past results of drawing balls from the urn will not help you much.
It’s the same thing with the stock market.
Econometrics study that exactly. The stock market in different time series. But is it worth it? The market of 1990 is so different than what it is now. Can we really expect similar outcomes?
Chapter 7: The Problem of Induction
The problem of induction is the following:
No amount of observations of white swans can allow the inference that all swans are white, but the observation of a single black swan is sufficient to refute that conclusion.
Empiricism (observing events and making assumptions about them) is dangerous.
If we look at data about car accidents, we will see that they mostly happen closer to people’s homes.
An empiricist could then tell you “you are more likely to have an accident in your neighborhood”. But is it the case? Or is it that people simply spend more time driving around their homes?
Data in fact, should not be used to prove a statement, but to disprove it. It’s not because something “has never happened” that it “never happens”.
Consider these two statements:
- A: no swans are black because all swans I have seen were white.
- B: not all swans are white
You cannot make statement A. But one black swan would be enough to make statement B.
Another logical flaw in predicting the future based on the past is if there were unexpected events in the past. If the past did not resemble its own past, why would our future resemble our past?
Furthermore, markets, such as life, do not have symmetric returns. It doesn’t matter how many times you win if only one loss is bigger than all of your gains.
Maximizing your chances of winning only does not lead to maximizing the total expectation of total gains.
Karl Popper came up with a great answer to the problem of induction. According to him, there are two types of scientific theories.
- Theories that have been tested and proven wrong (called falsified).
- Theories that have not yet been proven wrong.
Why is the theory never right? Because we will never know if all swans are white.
In a way, science is something we can falsify. If we can’t, it’s not science.
-> science should not be taken as seriously as we think.
Newtonian physics was disproved by Einstein. This is why Newtonian physics is science.
Astrology is not science, because the astrologist, upon explaining his mistake, could always say “well, Mars may have been slightly off, which would explain that…”
Karl Popper did not think that information could increase your knowledge, at least for certain disciplines (we don’t know which ones). The reason is that knowledge and discovery are not so much about what we know, but about what we don’t know.
To quote Popper:
These are men with bold ideas, but highly critical of their own ideas; they try to find whether their ideas are right by trying first to find whether they are not perhaps wrong. They work with bold conjectures and severe attempts at refuting their own conjectures.
For Popper, a theory cannot be verified. Only dismissed.
Our memory is not made to remember a list of random facts. We need a story, some logic that links events to each other. Our brain takes random facts and finds a link to explain them.
The induction problem works the same way. It is merely taking a bunch of particulars and making them general.
Doing so, we reduce the existing randomness.
Pascal said that it’s much better to believe in God than not because if God exists, the believer is rewarded. If he does not, the believer loses nothing.
Likewise, you can use statistics when it gives you an edge but refrain from doing so when it does not help you.
Part II: Monkeys on Typewriters
If you take an infinity of monkeys and give them typewriters, one of them would eventually come up with the exact version of the Iliad.
But would that same monkey also type the Odyssey?
Likely no. Yet, people often assume that in life.
The mainstream believes that if someone did well in the past, he’s likely to do well in the future.
It may actually be true, but it depends on two variables: how random was the activity he did well in, and how many monkeys were involved in the operation.
-> the greater the number of business people, the more likely one will do particularly well.
The problem is that you rarely can count all of them at the beginning. Those that lose usually vanish, and society only sees the winners, praising “intelligence” while it may just have been luck.
Chapter 8: Too Many Millionaires Next Door
The author tells the story of a couple earning 500k per year, living in Park Avenue in New York. Despite the fact they did better than 99.5% of people in the country, they consider themselves as losers. Why? Because they compared themselves to their neighbors, all successful people.
No losers live on Park Avenue.
There is a survivorship bias. They only see those that succeed. They don’t see the people that really failed.
Such a problem is found in the book The Millionaire Next Door.
The book explained how people that became millionaires became so by simply…spending as little as possible.
There are two problems with these observations:
- Visibility Winners: the people selected were the winning monkeys.
- Bull Market: the period during which these people became millionaires was a period of extreme and rapid stock market positive performances.
Survivorship bias is chronic. Why?
Because we draw conclusions on what we see, not on what we don’t see. Because of that, we assume that what we see will be the most probable.
The survivorship bias implies that the highest performing realization will be the most visible. But it’s wrong. It’s wrong because the losers do not show up.
Some financial gurus, for example, “found out” that failing money managers in point A in time had better financial performances in point B than successful managers.
The only problem is that they made the study based on failing managers that had survived up to then. They didn’t take into account the failing managers that went bankrupt.
The same can be said of optimists. Optimists take more risks. Those who win show up and give the advice to be optimistic.
Those who lose disappear, and never give any advice.
Chapter 9: It Is Easier to Buy and Sell Than Fry an Egg
If you go to your dentist, the chances that he can take care of teeth only out of luck is…rather small. The same can be said with a musician. These people’s track record means something. There isn’t much randomness.
But not in business. A track record means little. Too much randomness.
Fooled by Numbers
Let’s take 10 000 fund managers and put them in a Monte Carlo generator.
They each have a 50% of chance to make $10k per year or lose $10k.
Those that lose $10k are done forever, we will never see them again.
Here’s how it will go, for five years.
| Year 1 | Out of 10 000, 5000 managers survive. 5000 change job. |
| Year 2 | Out of 5000, 2500 managers survive. 2500 change job. |
| Year 3 | Out of 2 500, 1250 managers survive. 1250 change job. |
| Year 4 | Out of 1250, 625 managers survive. 625 change job. |
| Year 5 | Out of 625, 313 managers survive. 313 change job. |
In the real world, newspapers and biographers will highlight some imaginary qualities that these managers had that “made them succeed”, blinded by the fact that they survived purely out of luck.
Let’s now take 10 000 incompetent managers.
Instead of 50-50, they have a 45% chance to make $10k, and a 55% chance to lose $10k.
After 5 years, we’ll still have 184 managers surviving. Less than 2% of the total, sure, but give them business suits and they’ll be regarded as “legendary investors”.
-> a huge population of bad managers will produce a small number of great track records.
-> the number of managers that survive depends on the number of managers that started.
The same thing goes for individuals experiencing “lucky streaks”. With a sample population big enough, at least one subject will deviate from the mean.
-> the larger the deviation from the mean, the more likely it is coming from luck rather than skills.
Life Is Coincidental
The author makes a range of observations of coincidences that are in fact the result of randomness.
From meeting people you know in random places to having two people in the same room with shared birthdays, these are just the consequence of a sample big enough for such a “coincidence” to happen.
A scientist once looked at data from roulette in a casino, and found a pattern in the result.
Even randomness does not look random.
Chapter 10: Loser Takes All—on The Nonlinearities Of Life
Nonlinearity results from a linear force exerted on an object that causes a disproportionate result.
Eg: you add drops of water into a vase, until one drop makes the vase overflow.
All the drops were the same, but one created a disproportionate event.
These dynamics are often called “chaos theory”. Chaos theory concerns itself primarily with functions in which a small input can lead to a disproportionate response.
They have in essence, nothing to do with chaos, but with asymmetry.
“A butterfly flaps its wings in India and there is a storm in the US” is an example.
The final outcome is often undeserved. The keyboard QWERTY, for example, wasn’t invented to help people type faster but slower, to avoid typewriters breaking down. Now that people learned to use them, we can’t change them.
This effect is called path-dependent outcome.
Microsoft is one of these outcomes. It was not the best software, and Bill Gates was not the best or smartest businessman, but it’s the one that won due to network effects. And when he won, he won big.
These effects go against economics, where the best is supposed to win.
Mathematics Inside and Outside the Real World
Our brain does not like what is non-linear. Progress is one of these things. You practice for a year, and suddenly, you get some results.
Progression isn’t linear. This is why most people give up in the first place. Those who go the extra mile are rewarded.
When It Rains, It Pours
Non-linearity is everywhere.
- Everyone wants to publish your book, or no one does
- You whether become really rich, or you stay poor
- It’s better to have 10 people that love your work than 1000 that appreciate it
The Internet has worsened this. Winner takes all.
Chapter 11: Randomness And Our Mind: We Are Probability Blind
Paris, or the Bahamas?
This chapter is about heuristics described in Daniel Kahneman’s book “Thinking fast and slow” (the book was written after Fooled by Randomness, so it’s normal Taleb speaks about it).
Check out this Elon Musk’s tweet containing 50 of them if you’re interested.
Why We Don’t Marry the First Date
Our brains are made out of modules each tasked to solve a specific problem. You use different parts to look at art or solve an equation.
On Causality
Causality is interesting when the magnitude of the effect is important. If you cycle from Paris to Shanghai with a friend, and beat him by one second, you can’t possibly say you’re better than him as randomness likely played a role.
If you beat your friend by one week, then yes, you can.
You also need to select the right variable that could have caused you to win when looking for an explanation. Is it sleep, diet, or simply genetics? Causality is complex.
Little changes or differences are often noise. Don’t mind them.
Big changes though, aren’t.
Part III: Wax in my Ears
When Odysseus went back to Athens, he had to pass by the mermaids.
They sang to attract the sailors that would throw themselves into the sea, then be killed. Odysseus wanted to listen to their songs, so he put some wax in the ears of the sailor to protect them and tied himself to the mast, instructing his sailors to never release him until they had passed the mermaids.
Odysseus is a hero. We aren’t. We can’t fight our emotions. So we should also put some wax in our ears.
Wittgenstein’s Ruler
Unless you have confidence in the ruler’s reliability, if you use a ruler to measure a table you may also be using the table to measure the ruler.
Unless a reviewer is qualified to review a book, a book review says more about the reviewer than about the book.
Chapter 12: Gamblers’ Ticks And Pigeons In A Box
The author explains that he caught himself several times repeating a routine that he practiced one day he made a lot of profit.
He was disappointed in himself. While he knew superstition to be pointless, his emotions thought otherwise.
Despite mainstream thinking, superstitions aren’t cultural, but biological.
The psychologist B. F. Skinner showed how both humans and animals constantly look for logical patterns in randomness to help them get what they want.
This is something that we need to deal with, as this is emotional and we cannot split ourselves from our emotions.
Chapter 13: Carneades Comes To Rome: On Probability And Skepticism
For the Greeks, nothing was written in stone. They often changed their mind and were probabilistically oriented, not dogmatic about their positions.
Explained simply, they didn’t mind doubting and changed their mind often!
Christianity in the middle ages was much more dogmatic, until the Renaissance. Today, society considers that changing your mind is bad.
Yet one of the most famous and successful people to change their mind quickly is George Soros. He is not path-dependent. He doesn’t take action in light of his previous ones. Each day is a clean slate.
The rest of the people are not like George Soros. We have a hard time letting go of an idea when we have invested time into it.
Changing your mind all the time is a rational (not an emotional) thing. People that lack a disposition to attachment in their brain often change their minds – they are called psychopaths.
Computing Instead of Thinking
A hedge fund almost bankrupted the stock market in 1998 when they made decisions based on their own certainty that they could predict the market.
When the crash happened, they didn’t blame it on them. They blamed it on the crash, and on their peers that they claimed attacked them.
When we succeed, we think it is thanks to ourselves. When we fail, we think it is because of randomness.
Chapter 14: Bacchus Abandons Antony
No matter how good we are at choosing, we cannot avoid either randomness or emotions. But we can do the right thing, take the heroic or dignified path, to get even with randomness – we can make things right when randomness shackles them.
Remain worthy in times of uncertainty.
Your behavior is the only thing in your life that isn’t random.
Postscript
First Thought: The Inverse Skills Problem
The higher you go on the corporate ladder, the more money you make. However, the higher you go, the lower the evidence you make any significant contribution.
This is the inverse rule.
There are skills that are visible (being a cook, a dentist, etc) and skills that depend more on randomness (like marketing or business).
The only way you can verify these skills is through repetitiveness. If the cook nails the steak a hundred times, he is skilled.
Consider the difference between judging on the process, and judging on results. People at the bottom of the hierarchical pyramid will be judged on both processes and results. The CEO though, will be judged on results alone.
However, the link between the skills of the CEO and the results of the company is tenuous. He may just need to be charismatic enough to get the company going. The environment plays a huge role in business, much less than in the case of the dentist, or the cook.
CEOs have nothing to lose. When they win, they win big, and when they lose, the shareholders lose.
Imagine two incompetent CEOs. They both make a random decision. One is a disaster and the CEO is fired. The other is a success and the CEO becomes wealthy.
The second CEO becomes a hero, despite being as incompetent as the first one.
When we evaluate CEOs, what we see isn’t necessarily what there is.
Second Thought: On Some Additional Benefits Of Randomness
When you lack knowledge, it enables you to be more relaxed. Those that seek to maximize pleasure are much happier than those who seek to be efficient at all costs.
Man is not made to have a schedule. We are not made to operate within constraints. Knowing everything isn’t always good.
When you optimize too much and notice a little bit of fluctuation (noise), this can cascade into panic.
When you allow for a little bit of randomness, it gives you a buffer, a tolerance threshold for the unexpected.
Unpredictability can be a strong deterrent.
Third Thought: Standing On One Leg
“We favor the visible, the embedded, the personal, the narrated, and the tangible; we scorn the abstract. Everything good (aesthetics, ethics) and wrong (Fooled by Randomness) with us seems to flow from it.”
For more summaries, head to auresnotes.com.
Did you enjoy the summary? Get the book here!
Subscribe to my monthly newsletter and I'll send you a list of the articles I wrote during the previous month + insights from the books I am reading + a short bullet list of savvy facts that will expand your mind. I keep the whole thing under three minutes.
How does that sound?
