What breaks a thesis

The silence around selling

The literature of quality investing is a literature of entrances. Moats, runways, management, price — thousands of pages on what deserves capital, and a slogan where the exit should be: our favorite holding period is forever. It is a good slogan. It is also not a policy. Forever is the intended outcome of a correct thesis, not a substitute for knowing what an incorrect one looks like.

The silence has a cause. Selling is where the two most expensive mistakes in long-duration investing live, and they pull in opposite directions. Selling a great business early is the most costly mistake — the loss is capped at nothing and the forgone compounding is unbounded; everyone who sold a generational winner in year three carries the arithmetic privately. Holding a broken thesis is the most common mistake — endowed, ego-protected, and disguised as patience. A selling discipline has one job: distinguish these two situations while they still look identical.

What does not break a thesis

Start with the clutter. The price fell: not a reason — volatility is the admission fee, and if the thesis was priced with room to be wrong, a lower quote is an improvement in the offer, not a deterioration in the business. The price rose: not a reason — trimming winners because they have grown is how portfolios systematically transfer capital from their best judgments to their mediocre ones. The macro darkened: not a reason to sell a business — the weather changes the posture of the edge, never the contents of the core; the moment a regime reading starts liquidating compounders, the framework has been abandoned. Someone smart disagrees, the position has gotten boring, the gain would look good realized: temperament, all of it, wearing analysis as a costume.

What remains after the clutter is short. Three reasons, and only three.

The thesis broke

Every position is a claim about a specific engine: this advantage, deepening for this reason, funding reinvestment at these rates. The thesis breaks when the engine stops — and the essential discipline is that this registers in trajectory long before it registers in earnings. The moat stops moving. The propulsive engine stalls while the defensive one still holds, and the business begins its slow conversion into a well-defended melting asset. Current numbers stay respectable throughout; they are the camouflage, not the evidence.

This is why the reasons for owning must be written down at entry — specifically, falsifiably, before money makes the mind motivated. What would this thesis look like when wrong? Which measure of trajectory, deteriorating for how long, retires it? A thesis memo with kill criteria converts the future sell decision into a decision already made — made on the day you could still think clearly, executed on the day you cannot. Without it, every disconfirming quarter becomes “noise,” every erosion “temporary,” and the holding period becomes forever in the worst sense.

One asymmetry deserves respect here: a broken thesis and a paused one look alike for quarters at a time. The tiebreaker is the substrate. A great business having a bad year operates in a growing system with its engines intact. A broken one has lost the system itself — the market shrinking, the surface area contracting — however intact the quarter looks. Sell the second. Endure the first.

The capital has a better claim

The second honest reason is opportunity cost, and it carries a deliberately high bar. Not “something else looks attractive” — something else offers materially better compounding, at equal or lower risk of permanent loss, by enough to pay for the frictions: the taxes, the spread, and above all the risk of being wrong twice — wrong to leave, wrong to arrive. Switching resets a judgment that time has already validated for one that time has not yet tested. Most swaps that clear a casual bar fail an honest one. But capital is finite and conviction is comparative; when the gap is wide enough to survive all three frictions, refusing to act is its own failure of discipline.

The portfolio cannot survive being wrong

The third reason is the province of risk, not conviction. A position can be thriving — thesis intact, trajectory rising — and still grow into a size where one unforeseeable error would breach the only inviolable rule: never an impairment deep enough to interrupt compounding. Concentration is not risk; concentration without a survivable worst case is. When success pushes a position past the point where the portfolio can absorb its failure, trimming is not doubt. It is the same discipline that sized it in the first place, applied to a larger number.

The default remains

None of this disturbs the default, and the default is inaction. Most quarters, for most holdings, the correct decision is nothing — the engines are running, the trajectory holds, and interruption has a permanent cost that patience does not. The three reasons are not an invitation to activity. They are the perimeter around the patience: the pre-agreed conditions under which holding stops being discipline and becomes its opposite.

Patience without a breaking point is not conviction. It is captivity — and the difference is decided before the position is bought.