When to sell a stock: the four rules that actually decide it

Buying is the decision people research. Selling is the decision that determines the return, and it is almost always made in the moment, under pressure, and rationalised afterwards.

The fix is not a better indicator. It is deciding the exit before the entry, when you have no position and therefore no opinion.

Rule 1: the thesis is finished

You bought for a reason. Write it down at the time, in one sentence, in a form that can be false. "The earnings recovery is not priced in" is a thesis. "It looks cheap" is not.

Sell when the sentence stops being true — including when it comes true. A thesis that has played out is as much a reason to exit as one that has failed, and it is the exit people miss, because the position is profitable and profitable positions feel like they are working.

Rule 2: the time is up

A seasonal or event-driven position has an expiry built into it. If you entered for a window that runs to mid-October, you exit in mid-October regardless of how it is going.

This rule feels arbitrary and is the most valuable one on the list. The alternative — holding a window trade past its window because it is up — converts a tested edge into an untested one. The statistics you relied on described the window. Outside it you have nothing.

Rule 3: the loss reaches the size you decided on

Set it before entering, as a currency amount you are willing to lose, and derive the position size from it rather than the other way round. Deciding "I'll risk 1% of the account on this" and then computing how many shares that allows is a completely different discipline from buying a round number of shares and wondering afterwards where the stop goes.

The number should be small enough that being wrong is boring.

Rule 4: something better needs the capital

The least discussed and the most common in practice. Capital in a position that is going nowhere is capital not in the next one. If your process generates more candidates than you have slots, the weakest holding is a sell by definition, even if nothing about it has gone wrong.

This only works if you actually rank candidates. Without a ranking it becomes an excuse to churn.

What is not on the list

Round numbers. A stock does not know what you paid for it, and "I'll sell when it gets back to breakeven" is the single most expensive sentence in retail investing.

How you feel about the last three sessions. Short-term price movement is mostly noise, and noise is precisely what a pre-committed rule exists to filter.

A number someone posted. Their entry, horizon, size and tax situation are all different from yours, and all four change the answer.

Set the exit date before you enter

1. 1Pick the ticker and open Trading Analysis.
2. 2Set the period to the window you are entering, and note the historical exit date — that is rule 2, decided in advance.
3. 3Check the worst year for that window. If you could not hold through that, the position is too big.
4. 4Use Discover to see what else is opening in the same period. That is rule 4: the ranking that makes a swap rational rather than restless.

Open TradeSeasons

Writing it down

Whatever the rules are, they have to exist somewhere other than your head before the position is open. A position with a written exit is a trade. A position without one is a hope with money attached.