Disposition effect

Consider the asymmetry in how the two halves of a round trip get taught. Thousands of pages exist on what to buy: screening, valuation, moats, quality, the margin of safety. Ask the same literature when to sell and it goes quiet, offers a platitude ("sell when the fundamentals deteriorate"), or worse, hands you a folk saying that is actively wrong ("no one ever went broke taking a profit," see selling compounders too early). This is strange, because a portfolio's return is determined equally by its exits and its entries, and the sell decision is demonstrably the one where skilled investors bleed the most value. There are four structural reasons selling is harder, and naming them is the start of the cure.

Why is selling harder than buying?

First, you own it, and so the whole behavioral catalogue now binds. At the buy decision you are a neutral outsider comparing a price to a value. The moment you own the position, M6.04's demons attach: loss aversion makes crystallizing a loss feel roughly twice as painful as the equivalent gain feels good, so you defer selling losers; the endowment effect makes you value the thing more simply because it is yours, so you hold what you would never buy today; sunk cost whispers that the money and research already spent are reasons to stay; and the disposition effect, the mechanical output of all of the above, makes you sell winners too early (to lock in the sure gain) and hold losers too long (to gamble against crystallizing the loss). None of these operate at the buy. All of them operate at the sell.

Selling is behavioral finance on hard mode.

Second, the cost basis contaminates the decision. The price you paid is a fact about your past and is irrelevant to the security's future, yet it installs itself as the reference point from which every subsequent move is felt as gain or loss (prospect theory). "I'll sell once it's back to what I paid," "I'm not selling at a loss," "it's a three-bagger, I should take something off": every one of these anchors the forward decision to a backward number. The market has no memory of your cost and owes you nothing. The single most important mental move in the whole discipline is to delete the cost basis from the sell decision and replace it with the fresh-money test.

Third, the sell decision is chronically under-instrumented. Almost every serious investor keeps some record of why they bought: a thesis, a note, a model. Almost no one records why they sold, or sets sell criteria in advance. So buys get the full apparatus of System-2 analysis while sells are made in the moment, under the influence of the position, on a System-1 impulse ("it's up a lot," "I can't watch this fall further"). The result is that your worst-instrumented decisions are also your most behaviorally compromised ones. That is a guarantee of leakage. Half the job here is simply to give the sell decision the same written, pre-committed structure the buy already has.

Fourth, selling is where the two error types diverge and most people invert them. There are two ways to be wrong at the exit, and they are opposites: selling a winner too early (forfeiting the compounding that drives the whole portfolio) and holding a loser too long (letting a mistake or a broken thesis metastasize). The correct discipline is asymmetric: patient and slow with winners whose thesis is intact, ruthless and fast with losers that are mistakes or broken theses. The disposition effect makes untrained investors do the exact reverse: impatient with winners, patient with losers. Everything here is engineered to invert that inversion, to make you slow where you want to be fast and fast where you want to be slow.

Should the price move decide whether you sell?

A grounding principle before the framework: the sell decision is orthogonal to the price move. Whether a position is up or down is, by itself, never a reason to sell or hold. A price that has fallen with the thesis intact is often a gift (the Hunter's cue to add, in Freeman-Shor's five tribes); a price that has risen far past value is a warning (see selling above value); a price that has fallen because the thesis broke is an exit (see selling on a broken thesis). The move is an input to the analysis, not the decision. The decision keys off the five reasons, each of which is about the business, the value, the alternative, or your portfolio, never about your P&L on the line.

The disposition effect, operationalized: Freeman-Shor's five tribes

Before the five good reasons to sell, study the anatomy of doing it badly, because it is astonishingly consistent and it has been measured on real elite investors. Lee Freeman-Shor ran a fund-of-best-ideas at an asset manager from roughly 2006 to 2013. He allocated capital to 45 of the investors he judged the best in the world under a deliberate constraint: each could hold only their ten highest-conviction ideas, so every position was, by the manager's own account, a bet they believed in. He then watched what they actually did, trade by trade, and wrote it up in The Art of Execution (2015). The headline finding is disorienting. These elite managers were right, in the sense of a position that made money, only about half the time, roughly a 49% hit rate (illustrative of the book's data). Their best ideas lost money as often as not. And yet many of them made excellent returns. The difference between the ones who made money and the ones who lost it was almost entirely what they did after they bought, their execution of winners and losers, not their stock selection. This is the empirical heart of why selling deserves its own module: at the elite level, exit discipline mattered more than entry skill.

What are Freeman-Shor's five tribes?

Freeman-Shor sorted the behavior into five tribes. Learn them as named patterns; the app's tribe drill and half of your own self-diagnosis is exactly this recognition.

How they handled losers (a position that has fallen):

• The Rabbits, the losing pattern. Faced with a falling position, the Rabbit freezes: does nothing, neither adding nor selling, like a rabbit in the headlights. They hold the loser and hope, letting a small loss compound into a portfolio-wrecking one. This is the disposition effect's "ride the loser" half made flesh, powered by loss aversion and sunk cost. Rabbits were the single biggest destroyers of return in Freeman-Shor's data. The Rabbit has no plan for a loss because they never wrote one.
• The Assassins, a winning pattern. The Assassin kills losers ruthlessly and unemotionally, by a pre-set rule: a price stop (commonly a fall of 20–33% from cost), a time stop (if the thesis has not begun to play out within a set period, sell), or a thesis stop (a load-bearing pillar breaks). The rule is decided in the cold state and executed without renegotiation in the hot one. Assassins cap the left tail; no single position can take them down.
• The Hunters, a winning pattern. The Hunter plans to buy more as the price falls, deliberately deploying reserve capital into weakness, but only where the thesis is intact and the fall is Mr. Market's mood, not a broken pillar. The Hunter starts with a partial position precisely so they have ammunition for the drop; averaging down is a plan, not a panic. (Note the razor's edge here, developed in cutting losses on a stock: adding to a falling-but-intact thesis is the Hunter; adding to a falling-because-broken thesis is the Rabbit's cousin, throwing good money after bad. The thesis, not the price, tells you which.)

How did the tribes handle winners?

How they handled winners (a position that has risen):

• The Raiders, the losing pattern. The Raider snatches small profits, selling a winner after a modest gain (10–20%) to "lock it in" and feel the satisfaction of a realized win. This is the disposition effect's "cut the winner" half, powered by risk-aversion-in-gains and the fear of round-tripping. Raiders systematically amputate their compounders before the compounding happens, guaranteeing they can never own a big winner. As the great error of selling compounders too early shows, that is the same as guaranteeing mediocrity.
• The Connoisseurs, the winning pattern. The Connoisseur runs winners for years, tolerating volatility and resisting the urge to clip, occasionally top-slicing an oversized position (see trimming a position) but keeping the bulk working. They "savor" a great business like a fine wine and let a 3-bagger become a 30-bagger. The Connoisseurs made the bulk of Freeman-Shor's profits, because in a positively skewed return distribution the entire game is owning a few enormous winners for a long time.

How do the winning tribes invert the disposition effect?

The mapping onto M6.04 is exact and worth stating: the winning tribes are the inversion of the disposition effect. Where the untrained investor rides losers (Rabbit) and cuts winners (Raider), the disciplined investor cuts losers (Assassin) or adds to intact-thesis losers (Hunter) and runs winners (Connoisseur). Freeman-Shor's practical prescription is therefore a two-line rule: when a position falls, never be a Rabbit: be an Assassin or a Hunter, decided in advance; when a position rises, never be a Raider: be a Connoisseur. The five reasons are the analytical content that tells you which winning move applies; the tribes are the behavioral shape that content has to take.