Exiting Trades Too Early: What the Evidence Says About the Winner Side

By Sofia Harchich | Trading Psychologist & Behavioural Finance Writer| thewealthmirror.com

Exiting trades too early receives a fraction of the attention given to holding losers, and the imbalance is not justified by the arithmetic.

A loser held past invalidation costs a defined amount — the distance from the stop to wherever it was finally closed. A winner cut at a third of its target costs an undefined amount, indefinitely, on every occurrence, and it does so silently, because a trade closed in profit does not appear in any review as a problem.

Four findings, read for what each supports, and one correction that changes the intervention.

Finding one: the asymmetry is documented and it runs both ways

Supported: strongly.

The disposition effect describes a consistent tendency to realise gains too readily and defer losses. It has been observed across retail and professional populations and it is one of the more robust findings in this area.

The framing usually stops at the loser half, which is where the drama is. The gain half is the same mechanism operating in the other direction and it appears in the same data.

Prospect theory supplies the shape: the value function is concave in gains, so each additional unit of profit feels progressively less significant while the possibility of losing what is already showing feels increasingly costly. A position up 40 points does not feel twice as good as one up 20. The prospect of it returning to zero feels considerably worse than twice as bad.

Finding two: exiting trades too early avoids the reversal, not the smaller gain

Supported: as a direct implication, and it is the most useful of the four.

Exiting trades too early is rarely a decision to accept less. It is a decision to remove a specific future experience — watching an open profit disappear.

That distinction changes the target. A trader who believes they are impatient will work on patience and will make no progress, because patience was never the constraint. What they are avoiding is a particular kind of regret that only exists while a position is open, and it can be addressed structurally.

The evidence for this is behavioural rather than experimental, and it is easy to check on your own record: early exits cluster around the point where an open profit first begins retracing, not around any target, any time of day, or any level.

The exit is aimed at the retracement you can already picture, rather than at the profit.

Finding three: the reference point moves as the position moves

Supported: well, and it explains the timing.

The reference against which outcomes are evaluated is not fixed at entry. It updates, and it updates toward the position’s high. The same update is behind moving the stop to breakeven too early: once the high is the reference, a return to entry registers as a loss rather than as nothing.

A trade up 40 points that retraces to 30 is, by the arithmetic, a trade up 30. Experienced against a reference that has moved to 40, it is a loss of 10 — and loss aversion applies to that phantom loss with the same weighting it applies to a real one.

This is why the pull to exit intensifies precisely when a position is working. Every new high resets the reference upward and creates a new quantity that can be lost, so the discomfort grows with the profit rather than diminishing.

Nothing in that sequence is irrational once the reference point is understood. It is a well-behaved response to a badly chosen benchmark, and the benchmark was chosen by a platform display rather than by the trader.

Finding four: the standard remedy targets the wrong step

Not supported, and it is why “let your winners run” has helped so few people.

The instruction assumes the failure is at the decision — that the trader, at the moment of exit, weighs the options and chooses badly. Installing a change between a trigger and a behaviour runs detect, regulate, decide, execute, enforce, and let your winners run is aimed at the decide step.

The break is usually earlier. The retracement produces a physiological response, the response produces an urge to resolve, and the resolution is executed before any weighing occurs. A trader who has been told to let winners run experiences this as a failure of resolve, and adds a rule at step five, which reaches nothing.

The early exit is the instability loop firing on a winner, and it is reinforced precisely because it works: closing the trade relieves the feeling at once, which teaches the loop to fire sooner the next time open profit appears.

The correction that changes the intervention

Read together, the four findings point at one structural change rather than at more discipline.

Define the exit as an event before entry, on the same terms as the invalidation. Not a profit target chosen for its size, but an observable occurrence — closes below the 20-period average on the hourly, the level that produced the move fails. The exit then requires the same thing the entry required: noticing whether something happened.

This works for the same reason a pre-committed invalidation works. It moves the decision out of the moment when the reference point has drifted and the state is elevated, into a moment when neither is true.

Three supporting adjustments:

1. Change the display. Track the position against its invalidation and its exit condition rather than against unrealised profit, which is the field that keeps resetting the reference
2. Scale rather than close, if the pull is strong. A partial exit at a written level satisfies the urge to resolve without ending the position, and it is a legitimate compromise rather than a failure
3. Log the counterfactual. For ten early exits, record where the exit condition would have triggered. This is the only way to establish whether the early exits are costing anything, and for some traders they are not

The counterfactual log, in detail

Everything in this article reduces to one measurement, and it is worth setting out precisely because the wrong version of it produces nothing.

For every position closed in profit, record three fields. The price at which you exited. The price at which your written exit condition would have triggered. And whether that condition existed before entry or was constructed afterwards.

The third field is the one traders skip and the one that determines whether the exercise means anything. An exit condition written after the fact is a rationalisation with a number attached.

Ten trades produce a distribution, and it resolves in one of three ways.

Exits consistently well short of the condition. Exiting trades too early is running, and the gap is the measurement. Note that the size of the gap matters more than its frequency — a small consistent shortfall costs less than an occasional large one, and the two call for different responses.

Exits close to the condition. Nothing to fix. Some traders arrive here and discover that a problem they had assumed for years was a story rather than a pattern, which is a genuinely good outcome for a fortnight of logging.

No condition existed for most trades. The commonest result, and it reclassifies the problem entirely. This is not exiting trades too early; it is exiting without a defined exit, and the remedy is one sentence written before entry rather than anything about patience.

A fourth field is worth adding once the first three are running: the open profit at the moment of exit, as a proportion of the eventual move. Sorted across twenty trades, it usually shows a threshold — a level of unrealised gain beyond which exits cluster tightly — and that threshold is a considerably more useful number than any general resolution to hold longer.

One caution about reading it. The counterfactual is not a record of money left on the table, and treating it that way produces a trader who holds too long out of regret. It is a record of whether your exits are following your conditions, which is the only thing you control. A condition that triggers at a worse price than your early exit is a condition working correctly on that occasion, and a run of those means the conditions need revisiting rather than your discipline.

When exiting early is correct

The third point deserves separating out, because this article would be dishonest without it.

Not every early exit is an error. A strategy with a high win rate and modest targets is supposed to close positions quickly. A trader whose recovery speed is poor may be right to bank a gain rather than carry an open position into a state they know they handle badly. And market conditions genuinely change; an exit taken because volatility collapsed is a judgement, not a flinch.

The distinction is testable and it is the same test used elsewhere in this cluster. If your exits correlate with how the position feels — retracement, duration, size of open profit — the mechanism above is running. If they correlate with conditions — volatility, session, structure — you are adapting, and there is nothing here to fix.

Most traders are confident they know which. Ten logged counterfactuals frequently disagree, and they disagree in a specific direction: exits cluster around retracements more tightly than around any market variable.

What the pull is protecting

Underneath the mechanism sits something worth naming, because it is usually mistaken for weakness.

An open profit is provisional. It is not yours in any sense that matters until the position closes, and holding it means holding an unresolved claim on something good. That is a specific and genuinely uncomfortable state, and the discomfort is not greed’s opposite — it is the same intolerance of not-knowing that produces the errors on the losing side, pointed at a favourable outcome instead of an adverse one.

Which means exiting trades too early and holding losers too long are one difficulty rather than opposite faults requiring opposite corrections — sitting with an unresolved outcome — expressed in the two directions a position can move. Traders who resolve one often find the other moves with it, and traders who work on them as separate problems tend to make progress on neither.

Run it on your winners

Ten trades, two columns, no changes.

For every position closed in profit: where the written exit condition would have triggered, and whether one existed. That is it.

If no exit condition existed for most of them, the finding is specification rather than psychology, and the fix is one sentence written before entry. If they existed and were pre-empted, the finding is the retracement mechanism, and the fix is the display change plus a partial-exit rule.

For the mechanism this shares with its mirror image, read why traders hold losing positions; for the cluster hub, fear and greed in trading.

The pattern quiz identifies which emotional pattern your behaviour under pressure fits, and early exits appear most often in the controlled pattern, where they read as prudence rather than as a problem.

The column most journals leave out

One column, ten trades.

For every position closed in profit, write down where your exit condition would have triggered — and whether that condition existed before you entered.

Exiting trades too early is either a real cost with a measurable size or a story you have been carrying, and ten rows settle which. Most traders find the third result: that no condition existed, which makes this a specification problem with a one-sentence fix rather than a psychological one with none.

---

About the Author

Sofia Harchich is a Trading Psychologist and Behavioural Finance Writer with a Master’s in Psychology. She works at the intersection of Jungian shadow work, neuroscience, and market behaviour — helping traders understand the psychology driving their decisions, not just the strategy.

Read more at thewealthmirror.com/about