The Best Book For Understanding Luck

Fooled By Randomness

By Nassim Nicholas Taleb

Rating: 10/10

Best Line #1: Clearly risk taking is necessary for large success — but it is also necessary for failure.

Best Line #2: The only article Lady Fortuna has no control over is your behavior. Good luck.

I want to start by thanking Shane Parrish and Josh Wolfe for a marvelous interview on Shane’s podcast The Knowledge Project. Here’s a link to the interview. The episode was posted earlier this week while I was struggling to find a way to productively write about our featured book, Fooled By Randomness. This is a fabulous work by Nassim Nicholas Taleb and it goes deep to my core but I’ve struggled to explain how and why.

Then I listened to the podcast and the very first statement crystallized much of what I’ve used this book for. To set some context, the interviewee is the co-founder of Lux Capital, a venture firm. So he knows something about risk and uncertainty:

If you’re humble to the idea that luck matters so much, then it opens you up to maximizing that as cheaply as possible.

Maximizing luck. Yes. That’s it. That’s a great deal of what our author helps us do. By introducing the deep presence of randomness in all aspects of life, Taleb puts us on notice that we cannot underestimate the fates. Can we create our own luck? I don’t think so. At best, you can understand its nature, reduce its degree of influence (i.e., improve the odds), and shift the chance of it working in a positive rather than negative manner. That last part is as much about mindset as it is about math.

We’ll explore this further, much as I tried to do in yesterday’s article. And before that, I spend a lot of time exploring the ephemeral quality of certainty based on past history. I’m proud, but uncertain, of the effort there to convey the idea that what’s familiar about the future, as drawn to the past, is the mistaken view we hold of both. This week’s articles are below:

History Doesn’t Repeat. But Mistakes Do.

History Doesn’t Repeat. Mistakes Do. Part Two.

Choose Your Risks. Ignore Your Rewards.

What Is Randomness?

Definitions matter here. So as we dive in, let’s establish the foundational concept. Call it luck, call it randomness, call it probability. It’s very tempting to use these words interchangeably. So here’s clarity on the core idea. Luck and randomness are the same thing here. Luck is just an easier, more personable term. You can ascribe value to it (bad luck, good luck). It’s odd to say “bad randomness” as randomness is the general, non-individualized expression of the concept.

Then there is probability. This book is so unique for its time because Taleb expanded the idea to its fullest extent. As he explains,

There are many intellectual approaches to probability and risk — ‘probability’ means slightly different things to people in different disciplines. In this book it is tenaciously qualitative and literary as opposed to quantitative and ‘scientific’.

And …

In this book probability is principally a branch of applied skepticism, not an engineering discipline.

And finally …

Probability is not a mere computation of odds on the dice or more complicated variants; it is the acceptance of the lack of certainty in our knowledge and the development of methods for dealing with our ignorance.

Randomness Isn’t Your Spotify Shuffle

Again, these definitions matter. We have a hard time really understanding randomness. Our pattern-matching behavior, so deeply-ingrained in our minds, won’t let us go.

Consider Spotify. When it applied a genuine randomized algorithm to its shuffle function, people complained that the player was not actually random. Users detected patterns. And if there are patterns, one intuits, there can’t be randomness. Which is false but nonetheless easily assumed. As one engineer at Spotify told the BBC,

“Our brain is an excellent pattern-matching device. It will find patterns where there aren’t any.”

So they changed the algorithm. They made the random generator less random. So that it would feel more random to users.

Which is to say that we aren’t simply fooled by randomness. We are also fooled by non-randomness. Incredible.

A Monte Carlo In Our Mind

As written before, Taleb’s work has rich value not only for the writing he provides but the fantastic salad bar of concepts and ideas that come with it. I read this book a long time ago but, in doing so, I learned about Monte Carlo methods of prediction and problem solving. It is now my favorite method of approaching problems. Problem is, I don’t really have a monte carlo machine for anything outside an Excel spreadsheet.

The method is still useful as a means of understanding and leveraging randomness. As a brute force approach, it applies random values for defined variables in as many combinations as possible to distribute every outcome onto a range. This is something akin to the Infinite Monkeys idea. If you have two variables, monkeys and typewriters, and combin them in infinite ways so that infinite monkeys write on infinite of typewriters, one of these combinations of monkey and typewriter will produce Shakespeare’s Hamlet. Word for word. Guaranteed.

I don’t think there’s a computer out there that can give us a real sense of the probability for this but I’d say it’s … pretty low. All the same, it is a genuine outcome that could be produced by that combination of variables. I think it’s helpful to know this. It broadens our sense of possibility and gives us a new respect for the hilarity of randomness.

Informally, with our own thought experiments, we can consider a Monte Carlo approach for finding our lost keys.

If I lost my keys sometime in the past two days, I can take a map of the entire geographic area I have inhabited in those two days.

I can then add the amount of time I stayed in each portion of the area and frequency that I revisited those areas.

I can then generate a series of points to signify the potential location of my keys. More points would assigned to places I spent the most time.

And I can goof around with more variable. I might throw in a variable to account for “transfer stations”. For example, I’m more likely to grab my keys in moments where I am entering my vehicle from the garage. Or my home from the front door. Or from the moment that I hang my coat, transferring it to the closet, with my keys inside perhaps. These spaces get higher values based on the actions.

I can then combine these variables and map them by location in the full geographic area to produce values that range from very unlikely (low transfer activity, low amount of time spent in the given area, low frequency of revisiting the area) to highly likely (high transfer activity, high time spent, high frequency revisiting (e.g., the foyer closet where my coat is hung).

The Monte Carlo simulates these various combinations to optimize probability of containment (how likely are the keys here?) and probability of detection (how easily can I find the keys if they are here?). This tells me where to look first. And over time, the keys are eventually found.

Does anyone really think this way when looking for their keys? Sort of. But what’s interesting about the Monte Carlo approach is that it gives us a chance to understand broader possibilities. As explained by Taleb:

By dint of playing with my Monte Carlo engine for years I can no longer visualize a realized outcome without reference to the nonrealized ones.

Similarly, thinking in these ways gives me a better sense of the rich pageantry that is life. Of all the broad possibilities and unrealized outcomes, it is true that my keys could have absolutely fallen on the waterfront dock, been picked up by a pelican, and have since been stowed away in a nest overlooking the Pacific Ocean.

I think this is how the Coen Brothers make their movies. By imagining all those possible scenarios and strictly avoiding those that are most probable.

Anyway, the real point here is that a Monte Carlo engine is a formal dance with uncertainty, a chance to understand the key components of a given situation or problem and consider every way they could be combined to create outcomes both desired, undesired, and otherwise — without this work — unknown.

This reduces your risk of being surprised and broadens the thinking so that you don’t stay grooved inside the idea that you only lose you keys in your garage. Is that the most probable? Yes. But there are other probabilities that are just underneath the most-likely that are worth considering, too.

Does this sound exhausting? Sure! At first. But it is a form of mental exercise and, like all exercise, we can build endurance to do more and enjoy a higher level of fitness as a result.

More importantly, we can use this Monte Carlo model to reassess our variables. The more we understand the mechanics of a situation, the more we can trade certain variables for others and see much different results in the combinations. This gets to speculation of future options more so than searches for lost keys.

Crafting Our Bets. Not Just Placing Them.

So in terms of speculation, there’s no such thing as a sure thing. That’s for sure. But your Monte Carlo can find and capitalize on variables that have more ranges of positive possibility and less ranges of negative possibility. If your Monte Carlo shows a high number of combinations that are negative (a stock losing money, a potential job becoming stressful in year two), you should know that and reconsider your options. Can you combine a different variable that has less negative combinations? Can you price these options so that they are more acceptable?

This is easily understood in stock trading but I want to consider it in a more qualitative, vague sense with a job opportunity. So imagine a potential job that has great room for growth, great salary, but horrible culture and a lot of stress. What would such a job do to your health and happiness? You can answer that with a broad array of possible combinations from making you perfectly healthy and happy (less likely) to a little richer but not as happy or healthy anymore (more likely).

Based on those likelihoods, do you roll the dice?

Good question. But it’s not the only question. At this point, the job offer is merely defined by its initial presentation. This is the bet, the gamble, being offered to you. But it’s not the only bet available. A good Monte Carlo approach can show how there are other combinations, many of which could be more favorable.

The casino, this ain’t. In most things, you can not only choose to play the game or not play the game. You can negotiate the rules and payoffs and shape the game to your liking.

So before you decide to roll the dice, consider what happens if you change a variable or two (salary, expectations, workload, title, etc). Does this make for a better bet? A better gamble? What’s the range that works for both you and the job recruiter?

These questions open up broader possibilities and create a chance to build safer bets.

More importantly, it provides a range of possibilities within the bounds of randomness that works for you both, the hiree and the hirer.

This, I think, is a really important idea: building safer bets. We often think we are handed an opportunity and we need to either accept it or not. That the risks out there are solely crafted by The Fates and we must take them at the terms provided. Studying Taleb and others has taught me that we can build our own bets; we have the choice. And so long as our Monte Carlo generator itself is functioning properly, you can understand what a safe bet means. This echoes the quote from Josh Wolfe, which I introduced at the start of the review:

If you’re humble to the idea that luck matters so much, then it opens you up to maximizing that as cheaply as possible.

Maximizing luck means maximizing the bets we choose to make in a way that favor us to our total level of contentment. Win or lose, it’s on our own terms. This is a better dance with uncertainty.

The Score Takes Care of Itself

There is a line in Taleb’s book where he describes one of his former bosses, a man given the pseudonym Jean-Patrice:

He is one of the very rare people who have the guts to care only about the generator, entirely oblivious to the results.

It reminded me of a book title that I absolutely adore. The book is called The Score Takes Care of Itself by the legendary football coach Bill Walsh. The title says it all. In the book, Walsh describes his methods and approaches for coaching and leadership. And whether it’s Andy Grove book on management (last week’s review is here) or Walsh’s philosophy, what I see is a generator.

This is something I get cranked up about. I am often frustrated by the managers and leaders who scurry about chasing specific outcomes, wholly outside of their control, while ignoring the mechanics that produced the result. This gets us back to Thinking In Systems (book review here). But it also gets us to what I admire most about Taleb.

Consider this line:

At a given time in the market, the most successful traders are likely to be those that are best fit to the latest cycle.

It appears that Taleb’s former boss, Jean-Patrice, understood this quite well. Market cycles come and go. Sure, you can make it all about timing but time marches on. A proper generator, system, method, and philosophy will outlast them all. This is why a simple algorithm, performed with proper information, will outperform a person equipped with the same information in instances of regular, cyclical activity.

Because of randomness. And emotion. And perception. And a lot of other things, too.

The point? In instances of regularity, a’la management, it is far better to build a model that is right 51% of the time across cycles and circumstances than to live ad hoc in a given cycle only to find yourself without a foundation once that cycle passes. In other words, don’t chase fads. Don’t live and die by your own whims and creativity in the moment. And don’t take one bad result to mean the failure of an entire system.

Magnitude Matters More

That said, I think Taleb would smack the back of my head if I didn’t make a distinction between regularity and rarity. Results do matter. A lot. Specifically, the really bad ones that are possible if we don’t build an approach (i.e., a generator) that prevents us from total loss during a rare event. Consider this line:

It is not how likely an event is to happen that matters, it is how much is made when it happens that should be the consideration. How frequent the profit is irrelevant; it is the magnitude of the outcome that counts.

This gets to asymmetry. And here is where the properties of randomness are their most interesting and most, well, profitable. Consider the following:

The best description of my lifelong business in the market is ‘skewed bets’, that is, I try to benefit from rare events, events that do not tend to repeat themselves frequently, but, accordingly, present a large payoff when they occur.

Venture capitalists do the same thing. And I’ve come to really appreciate the art of this work after Josh Wolfe’s insights on the aforementioned podcast. I won’t spoil it but I should mention something really profound in the work. It has to do with the power of having “nothing to lose”.

“When we started Lux, people told me I was taking the biggest risk. But I looked at it and said, ‘What’s my risk? I come from no money, my mom is school teacher in Coney Island, Brooklyn. My worst case scenario is that I fail and I go back and join the mainstream.”

“When you start a company, everything’s a risk. Financing risk, technology risk, management risk, product risk. In fact, when I think of this, I think of the first law of Thermodynamics, that energy is not created nor destroyed. Risk and value themselves are just changing form. And every time you can kill a risk, subsequence value gets created.”

How do you deal with randomness and risk? How do you minimize it? This might be the best, most effective way: by having nothing to lose. Nothing to lose means no risk. Which means all value. Which means all upside. Which gets back to yesterday’s idea of upside risk. You can only go up. Or at worst, sideways.

Such an approach creates a natural skewness or asymmetry of the sort that Taleb writes about. I find it aesthetically beautiful and quite powerful, too.

Random Mistakes

That said, there are few instances where Taleb’s writing can be easily boiled down. This is what makes the work so hard here. I have to translate his deep narrative into digestible sound bites and he does not make it easy. Of course, my real intent here isn’t to capture his gestalt in a single article. I just try to communicate this easily in order to persuade you to buy his book.

In the interest of being more persuasive, then, (and helpful, too) I have a bullet list here that I derive from his text. These are largely his words but I offer them here for the person who is scrolling/scanning the page and is looking for the easy stuff.

There are many mistakes we make when we are fooled by randomness. Taleb mentions a few here:

What We Do When We’re Fooled By Randomness

Mistake 1: An overestimation of the accuracy of one’s beliefs. Without fuller testing, you can’t dispute the idea that what worked before could be as much coincidence as correctness.

Mistake 2: A tendency to get married to positions.

Mistake 3: A lack of a precise game plan ahead of time for what to do in the event of losses.

Mistake 4: An absence of critical thinking. A lack of stop losses (i.e. a predetermined exit point).

Mistake 5: Denial. No acceptance of what happens.

This ends my obligatory bullet list.

A Strong Voice. An Acquired Taste. An Important Work.

My problem is that I am not rational and I am extremely prone to drown in randomness and to incur emotional torture.

Say what you will about Taleb. At least he has an informed and passionate point of view. There is a bedrock to his logic that is often clear, ancient, and sound. If he could treat Twitter with the distaste he treats other modern (i.e., short-lived, unproven) things, maybe his bombastic nature wouldn’t get the better of him. His belligerent tweets are, well, belligerent. That said, he knows what he’s doing. I think Taleb always knows what he’s doing. And his willingness to be vulnerable in his foibles helps me be aware of my own. In this way, his books are so important and his voice is definitely needed. Warts and all.

I can’t recommend the book enough. I’m always careful to select the books that really matter. In this case, Taleb’s first work is the first instance of something that is more than an instructional text. Of all the books I’ve reviewed so far (this is #21), Fooled By Randomness is the first to give us a truly rigorous narrative. And it opens up so many doors for the curious mind. Here is where I first truly heard about inference, induction, falsifiability, and the very nature of knowledge. All of which is offered in a beautifully-opinionated scroll.

This topic is potent. Important. It has really helped me and I hope it can help you, too. As a last word, though, I urge caution. This book can open up a significant amount of thinking and ideas that you probably won’t be able to share with anyone else who hasn’t read anything of this nature. In fact, the knowledge of a book like this makes me lonely at times. This is the first moment that I’ve been able to share it.

Doing so has been a challenge. It’s hard not to go off the deep end. If you aren’t fooled by randomness, you might be swallowed by it.

Here’s a link to the book on Amazon.

Principles and Mental Models

• Probability is not a mere computation of odds on the dice or more complicated variants; it is the acceptance of the lack of certainty in our knowledge and the development of methods for dealing with our ignorance.
• Bad information is worse than no information at all.
• That which came with the help of luck could be taken away by luck. Things that come with little help from luck are more resistant to randomness.
• Skewness: it does not matter how frequently something succeeds if failure is too costly to bear.
• Mild success can be explainable by skills and labor. Wild success is attributable to variance.
• Do not judge a performance in any given field by the results. Just it by the costs of th elatnerative.
• The quality of a decision cannot be solely judged based on its outcome.
• Denigration of history: the thought that the sorts of things that happen to others will not happen to you.
• Care only about the generator; not the results.
• Rational thinking has little, very little, to do with risk avoidance.
• Epiphenomenalism: the illusion of cause-and-effect. For example, risk management has less to do with actual risk reduction than it has to do with the impression of risk reduction.
• Random is not equiprobable. Some outcomes will give higher probability than others.
• Prepare for all outcomes. Always prepare.
• Stop loss: a predetermined point of exit.
• It is not how likely an event is to happen that matters, it is how much is made when it happens that should be the consideration. How frequent the profit is irrelevant; it is the magnitude of the outcome that counts.
• Rare events are, by nature, undervalued.
• The Lucas Critique
• Falsifiability
• An open society is one in which no permanent truth is held to exist. A good model for society that cannot be left open to falsification is totalitarian.
• Survivorship bias: the history performing realization will be the most visible.
• No one accepts randomness in their successes, only their failures.
• The overfitting effect of data analysis.
• Real randomness does not look random.
• Virtue is its own reward.
• Satisficing
• Bounded Rationality
• Don’t do to others what you don’t want them to do to you; the rest is just commentary.
• We favor the visible, the embedded, the personal, the narrated, and the tangible; we scorn the abstract.
• Economics does not describe how people do act but rather how they should act.