This governs one question you will answer many times a year, usually badly: when something worked, was it you? Taleb's claim is that people habitually read the left column of his table of confusion as the right one, luck as skill, noise as signal, an anecdote as knowledge, and that the trained specialist is often the worst offender because he trusts his own expertise. A large part of the book is about what to believe regarding your own track record rather than what to do on Tuesday, so those items sit below under Hold.

Do

Judge a decision by the process that generated it, not by the one result you can see. Taleb contrasts two bosses: Kenny, who praised traders on their recent numbers and could not grasp survivorship bias, and Jean-Patrice, who cared only about the quality of the risk-generating process and was indifferent to short-term results. Kenny's traders eventually blew up. [Source: "Chapter Two: A Bizarre Accounting Method"]

Ask how many people were playing before you credit the winner. Simulate 10,000 managers in a fair coin-flip game, cutting anyone with a losing year, and roughly 313 finish with five straight winning years on chance alone. Run the same simulation with managers rigged to lose on average and about 184 still emerge with a glittering five-year record. With ten competitors that streak means something; with ten thousand it means nothing. [Source: "Chapter Nine: It Is Easier to Buy and Sell Than Fry an Egg"]

Reconstruct the losers before you draw a lesson from the winners. To test whether winners keep winning, you have to start from the full population at a fixed past date and track everyone forward, including the ones who dropped out. A backtest of a contrarian strategy that only included managers still in business at the end of the period silently deleted every manager who was shut down, which invalidates the whole comparison. [Source: "Chapter Eight: Too Many Millionaires Next Door"]

Look at your results far less often. For a portfolio with a 15% expected excess return and 10% volatility, the ratio of noise to real performance runs about 0.7 to 1 over a year and roughly 1,796 to 1 over a second. Since losses hurt more than equivalent gains please, checking every minute produces enormous net emotional pain and no additional information about the same strategy. [Source: "Chapter Three: A Mathematical Meditation on History"]

Size the outcomes before you count how often you are right. A bet with a 999-in-1,000 chance of winning $1 and a 1-in-1,000 chance of losing $10,000 has a negative expected value near $9 while winning almost every time. This is why Taleb refuses "bullish" and "bearish" as words: he could think a market was more likely to rise and still be positioned short, because the downside was larger than the upside. [Source: "Chapter Six: Skewness and Asymmetry"]

Use history to find opportunities, never to set your risk limit. Taleb's personal rule, taken from Popper's asymmetry, is that data can disprove a proposition but never prove one, so he lets inductive methods make aggressive bets and refuses to let them bound his exposure. Carlos and John used the same historical data both to place their bets and to measure their safety, and were wiped out. [Source: "Chapter Seven: The Problem of Induction"]

Test whether you are still holding a position or just married to it. The test is one question: would you buy this today, at this price, if you did not already own it? If the answer is no and you are still holding, you have path dependence of belief rather than a live assessment. Taleb's model here is Soros reversing a heavily bearish position within days once new evidence appeared, with no embarrassment at all. [Source: "Chapter Thirteen: Carneades Comes to Rome: On Probability and Skepticism"]

Build structural avoidance instead of relying on willpower. Taleb denies himself access to his own performance reports unless results cross a predetermined statistical threshold, and compares it to keeping no chocolate under the desk. He identifies with Odysseus's sailors, who plugged their ears with wax, not with Odysseus, who was strong enough to listen tied to a mast. [Source: "Chapter Twelve: Gamblers' Ticks and Pigeons in a Box"]

Fix your conduct so the result cannot move it. Dress and comport yourself well on your worst day, stay courteous to subordinates when you have lost money, avoid self-pity and blame, and treat victory and defeat as the same occasion for dignified behavior. Taleb's argument is narrowly practical: your behavior is the only article over which fortune has no control. [Source: "Chapter Fourteen: Bacchus Abandons Antony"]

Don't

Don't read visible wealth as evidence of a sound process. Nero, a conservative options trader whose rule was never to risk more than a fixed amount regardless of probability, spent years envying his neighbor John's bigger house and sports cars. John was fired in September 1998 when a leveraged interest-rate-spread bet collapsed. Profitable people also develop a confident posture and manner through serotonin whether the profit came from skill or luck, so behavior gives you nothing either. [Source: "Chapter One: If You're So Rich Why Aren't You So Smart?"]

Don't discard the extreme values as outliers. Trimming extremes is a reasonable habit borrowed from fields like education where the tails do not matter much. In finance the rare extreme is exactly what does the damage, and the "peso problem" names the pattern: long stretches of apparent calm followed by a sudden severe break, with statistical confidence about an event's absence growing far more slowly than confidence about its presence. [Source: "Chapter Six: Skewness and Asymmetry"]

Don't rename yourself a long-term investor when a position moves against you. Taleb lists what Carlos and John had in common: overconfidence in their own model of value, emotional marriage to positions instead of testing them, changing their story from trader to investor once losing, no predetermined exit plan, and dismissing the price on the screen as mere liquidation. Carlos lost roughly $300 million in one summer after $80 million in cumulative prior gains. [Source: "Chapter Five: Survival of the Least Fit - Can Evolution Be Fooled by Randomness?"]

Don't mistake fluent, impressive-sounding language for content. Fed with the right vocabulary, a text generator produces grammatically perfect postmodern academic prose that means nothing, and a random recombination of stock business phrases produces a speech indistinguishable from a real CEO's. Genuine scientific argument cannot be assembled by chance recombination in the same way, which makes this a usable test. [Source: "Chapter Four: Randomness, Nonsense, and the Scientific Intellectual"]

Don't treat current dominance as proof of underlying merit. QWERTY was designed to slow typing down and locked in anyway once enough people were trained on it. People use widely-used software largely because other people use it. Taleb's model for this is the Polya urn, where drawing a color raises the chance of drawing it again, so early luck compounds rather than averaging out the way a fair coin does. [Source: "Chapter Ten: Loser Takes All - On the Nonlinearities of Life"]

Don't reason from the single most likely outcome. Asked to imagine a 50/50 vacation, the mind produces Paris or the Bahamas, never a blend. Taleb's own cancer diagnosis carried a 72% five-year survival rate and registered emotionally as pure hope rather than as a real 28%. The trading version is option blindness, defaulting to the most likely outcome rather than the probability-weighted one, which is how premium sellers collect small steady gains until one rare loss removes them. [Source: "Chapter Eleven: Randomness and Our Brain: We Are Probability Blind"]

Don't accept a "found" rule from a machine that searched until it found one. Data-mining software will always eventually turn up a historical trading rule that worked. The same structure runs the chain-letter scam that mails contradictory predictions to thousands of people and recontacts only the ones who got the right letter, and it runs scientific publishing too, where null results are far less likely to be published than significant ones. [Source: "Chapter Nine: It Is Easier to Buy and Sell Than Fry an Egg"]

Don't assume rigor in one domain carries into the others. Nero survives cancer, gets rich honestly, stays alert to financial blowups, and then dies landing a helicopter alone on a windy day, having taken up flying out of frustration with London traffic. His probability-consciousness never once got applied to everyday physical risk. [Source: "Epilogue: Solon Told You So"]

Hold

The result you got is one of several that could have happened, and the others still count. Pull the trigger for $10 million on a six-chamber revolver with one bullet and five histories make you rich and admired while one produces an obituary; the accountant sees only the money. Holding this lets you refuse money that was earned dangerously and refuse to feel foolish about it. Refusing it means the only evidence you will ever accept is the branch you happened to land on. [Source: "Chapter Two: A Bizarre Accounting Method"]

Given enough time, both the lucky fool's advantage and the unlucky expert's deficit erode. Very long sample paths converge toward resembling each other regardless of starting luck. Held, this is a reason to keep operating rather than to conclude from a bad stretch that you are wrong. Not held, every short run becomes a verdict. [Source: "Chapter Three: A Mathematical Meditation on History"]

Nobody who has not yet met the rare event has been shown to survive it. Taleb rejects the popular Darwinism that says competition promotes the fittest, because fitness is relative to a specific and often short regime, not an unconditional property. An animal, or a trader, can look maximally fit while simply not having encountered the event its strategy was never equipped for, and the longer the streak lasts the more overconfident it becomes. [Source: "Chapter Five: Survival of the Least Fit - Can Evolution Be Fooled by Randomness?"]

Your mind manufactures causal connection whether or not any exists. Skinner's pigeons, fed on a random schedule unconnected to anything they did, developed elaborate personal rituals like head-swinging and spinning. Taleb caught himself repeating the route, entrance, and coffee-stained tie from a day he made unusually large profits. Holding this reclassifies your own hunches as data about your wiring, not about the world. [Source: "Chapter Twelve: Gamblers' Ticks and Pigeons in a Box"]

Certainty was never on offer, and changing your mind is not a defect. Carneades argued brilliantly for justice before the Roman Senate and then argued the opposite case with equal power the next day, specifically to demonstrate that certainty is unattainable, which so alarmed Cato that he had the ambassadors expelled. Held, reversal becomes cheap. Not held, every past statement is a debt you keep paying. [Source: "Chapter Thirteen: Carneades Comes to Rome: On Probability and Skepticism"]

No life can be called fortunate before it is over. This is Solon's warning to Croesus, which Taleb adopts as a personal motto and restates as the skewness problem: it does not matter how often something has succeeded if one failure can be catastrophic. Holding it keeps the current score provisional. Refusing it means treating whatever has not yet broken as proof that it cannot. [Source: "Epilogue: Solon Told You So"]

The one line

Stop asking whether the outcome was good and start asking how many ways it could have gone, how many people were playing, and whether you would still be standing in the versions you did not get.