Disposition Effect: Why Purchase Price Distorts Exits

Two journal entries from the same breakout method can look almost identical at the moment of entry, and the disposition effect is often what pulls them apart later. A stock clears a base pivot, the initial stop goes in, the plan says hold for several weeks while the trend stays intact. A few weeks later one position is up 5.9% and the other is under water. The first trader exits early because the gain feels fragile. The second keeps a position past the exit the plan already named, because closing it below the purchase price feels like admitting a mistake. Both decisions point back to the same number, the price paid.

This guide uses two invented trades to show how the bias can surface in growth-stock breakout trading, how to separate it from a legitimate planned exit, and why its cost has to be measured in your own records rather than assumed.

The disposition effect, defined narrowly

The disposition effect is the tendency to realize gains more readily than losses. Hersh Shefrin and Meir Statman gave it the name in a 1985 paper, and later studies of brokerage account records reported the same tilt among individual investors. The idea is simple. Positions showing a profit tend to get closed sooner, and positions showing a loss tend to be kept longer, at rates the investor’s own strategy doesn’t explain.

That last clause carries the definition. Taking a profit is ordinary, and so is holding a position that sits below cost. A trader who takes a planned partial profit at a target defined before entry is following a method. A trader who gives a position room inside a stop that was set on day one is also following a method. The bias describes something narrower: the purchase price, rather than the setup, deciding when the position ends.

The common misread is to label every early exit or every long-held loser as the disposition effect. From the outside, a disciplined exit and a biased one can produce the same trade record. The difference sits in the reason, and only a reason written down at the time can show it.

Why the purchase price becomes a reference point

Once a position is open, the entry price becomes the zero line against which every later price gets judged. A stock at 66.10 bought at 62.40 registers as “up 3.70.” The same stock at 66.10 bought at 69.00 registers as “down 2.90,” even though the chart, the trend, and the risk from here are identical for both holders. The price chart has no memory of what either of them paid.

Several explanations have been offered for why that zero line pulls on decisions. They are possible contributors, and none of them is a diagnosis that applies to every trade:

• Reference dependence. Outcomes are felt as gains or losses relative to a starting point, so the same price can feel like a win or a defeat depending on where you came in.
• Regret. Closing a loser makes the error final and visible, while a small open gain feels like something that could be taken away.
• Mental accounting. Each position gets its own little ledger, and closing a ledger in the red feels different from closing one in the black, even when the portfolio result is the same.
• Getting back to even. The purchase price turns into a target in its own right, so “out at breakeven” starts to replace whatever exit the setup called for.

I find the last one easiest to spot in my own notes, because it leaves a phrase behind. When a journal entry says “out when it gets back to 38.50” and 38.50 is the entry price, the exit is anchored to the purchase, and nothing on the chart points to that level.

Trade A: exiting an orderly advance because the gain feels fragile

Here is the first fictional trade. The declared method is a base breakout with a holding horizon of several weeks to a few months. The written plan, recorded before entry, reads:

• Entry 62.40 on a close above a 61.90 pivot, volume 1.8 times the 50-day average.
• Initial stop 58.70, giving a risk of 3.70 per share.
• Exit if the stop is hit, or on a weekly close below the 10-week moving average. No profit target.

Over the next three weeks the stock advances in an orderly way. It closes at 63.80, 64.95 and then 66.10 on the weekly chart, each pullback holding above the prior week’s low, with the 10-week line well below price. None of the exit conditions has fired. At 66.10 the trader closes the whole position. The note written at the time says: “Up almost 6%. Don’t want to watch this turn into a loss.”

That’s a gain of 3.70 per share against a risk of 3.70, or exactly 1R in the language of expectancy and R-multiples. The result itself tells you nothing about whether the exit was sound. What matters is the comparison between the note and the plan. The plan asked for a weekly close below the 10-week line. The note cites the size of the open profit and the fear of losing it. The setup hadn’t changed. The trader’s distance from the purchase price had. When I read a record like this, the first thing I check is which of the plan’s two exits the note mentions, and here it mentions neither the 58.70 stop nor the 10-week line.

Trade B: postponing a defined exit because the position is under water

The second fictional trade uses the same method and horizon. The written plan:

• Entry 38.50 on a close above a 38.10 pivot, volume 2.0 times the 50-day average.
• Initial stop 35.80, a risk of 2.70 per share.
• Exit if the stop is hit, or if price closes back inside the base on volume above average.

Eight sessions later the stock closes at 36.90, back inside the base, on volume 1.6 times the 50-day average. The second exit condition has fired. Two sessions after that it closes at 35.60, under the 35.80 stop, so the first condition has fired too. The trader moves the stop to 34.00. The note says: “Base still holding overall. Will exit when it gets back to 38.50.”

The plan treated a close back into the base on heavy volume as invalidation, which is the textbook warning sign of a failed breakout. The note offers a new reading of the base and a new exit, and the new exit is the purchase price. No fresh information appears in the note beyond the position being below cost. The risk has quietly grown from 2.70 to 4.50 per share, which is 1.67R measured against the original plan.

Whatever the stock does next is irrelevant to this assessment. If it recovers to 38.50 the following month, the decision still departed from the written plan for a reason the plan didn’t recognize. If it falls to 30.00, that later price is just as silent about the quality of the decision at 35.60.

Why later price outcomes can’t grade the decision

It’s tempting to settle the question by checking what happened afterwards. Trade A rose to 74.00, so the exit was a mistake; Trade B recovered, so holding was fine. That reasoning grades the decision by a price nobody could see when it was made.

A single outcome mixes the quality of the decision with the variance of the market. A sound exit can be followed by a rally, and a poor one can be rescued by luck. Hindsight also tends to rewrite the reason: a trader whose held loser recovered will often remember the hold as conviction, even when the note at the time said “back to even.”

The comparison that stays fair uses only what existed at the decision point: the written plan, the chart up to that bar, and the reason recorded then. That’s why the contemporaneous note carries so much weight in this kind of review. The structure of the note matters less than the fact that it was written before the outcome was known, which is the main point of trade journal design for this purpose.

When early profits and long holds are part of the strategy

Neither behavior is a bias in itself. Suppose the plan for Trade A had said: “Take one third off at 2R ahead of the earnings report, hold the rest on the 10-week line.” A partial exit at 69.80 (62.40 plus 2 times 3.70) before a scheduled report would then be the method working as written. A planned profit target, or an exit ahead of a known event, is a strategy choice with its own logic. The article on take-profit strategy covers the trade-offs among those choices.

Continued holding works the same way. If Trade B had pulled back to 37.20, below the 38.50 entry but above the 35.80 stop and without closing back inside the base on heavy volume, holding would be the plan doing its job. A position below cost and inside its original risk is behaving within the room it was given.

Two conditions separate these cases from the disposition effect. The reason was defined before entry, and the reason still applies at the moment of the decision. The pre-trade checklist is where many traders record that reason so that later, under pressure, there is something fixed to compare against.

New information can also justify a change. A guidance cut, a halt, or a sharp change in market conditions can make a plan stale, and revising it is reasonable. The test is whether the note names that information, or whether the only new fact it mentions is where price sits relative to the entry.

How the pattern can bend the payoff profile

The reason this bias matters to a breakout trader is arithmetic. Breakout methods typically depend on a minority of large winners covering a majority of small, capped losses, a relationship the piece on win rate versus payoff ratio works through in detail. The disposition effect pushes on both halves at once.

A hypothetical example shows the direction. Say a method, followed as written, produces a 40% win rate with an average winner of 2.5R and an average loser of 1.0R. Expectancy is 0.40 times 2.5 minus 0.60 times 1.0, which comes to +0.40R per trade. Now suppose winners are repeatedly shortened to an average of 1.2R and losers stretched to an average of 1.4R, with the win rate unchanged. Expectancy becomes 0.40 times 1.2 minus 0.60 times 1.4, or -0.36R.

That calculation is an illustration, and the real effect has to be measured, because the inputs rarely stay put. Closing winners early can raise the win rate, since some trades that would have reversed get banked first. Holding losers can turn a few of them into small gains or scratches. The net change depends on the method, the market, and how often the behavior actually occurs. A trader’s own records are the only place to find that number.

Jesse Livermore described the same tension in his own terms: hope keeping a losing position open past its exit, and fear closing a winning one before the trend has finished. The Jesse Livermore trading lessons profile shows how a trend trader of that era thought about letting the market decide the exit.

A retrospective comparison of reasons, invalidation and behavior

Here is how the two trades above look once the plan, the note and the action sit side by side. This is a way of reading the record after the fact. It doesn’t add a new rule or a new stop to the method.

• Trade A. Planned invalidation: stop at 58.70 or a weekly close below the 10-week line. Reason recorded at exit: “Up almost 6%, don’t want to watch this turn into a loss.” Actual behavior: full exit at 66.10 with no invalidation condition met. The stated reason refers to the open gain.
• Trade B. Planned invalidation: stop at 35.80 or a close back inside the base on above-average volume. Reason recorded: “Base still holding overall. Will exit when it gets back to 38.50.” Actual behavior: both conditions met, stop moved to 34.00. The stated reason refers to the purchase price.

Laid out like that, each record shows the same gap. The written plan named one exit, the behavior followed another, and the reason given points to the distance from cost. One pair of trades proves nothing about a trader. Across 30 or 50 closed positions, though, a recurring cluster of entries where gains are closed before invalidation and losses are held past it, with notes that refer to the entry price, is the kind of evidence worth taking seriously.

What this kind of review can’t tell you

An isolated trade can’t establish a behavioral pattern. Trade A alone could be an impulsive day, a reasonable read of something the plan missed, or a bias. Only repetition across many trades separates those possibilities, and even then the evidence is a tendency, never a verdict on each decision.

Different strategies have different exit logic. A swing method with fixed targets will close winners sooner than a position-trading method that trails a moving average, and comparing the two by holding time alone would misread both. The disposition effect is measured against the trader’s own declared method.

Transaction costs and new information can justify changes too. A position that has drifted into a tiny gain may not be worth the cost of managing, and a material news event can make the original plan irrelevant. Self-reported motives are also imperfect. A note written in the moment may leave out the real reason or offer a tidy one, so the written record is the best available evidence, and an imperfect window into intent.

Awareness doesn’t remove the bias either. Knowing the term doesn’t stop the purchase price from feeling important at 66.10 or at 35.60. What the term offers is a precise question to bring to your own records.

Letting the setup decide the exit

The purchase price is a fact about you. The chart doesn’t know it, and a breakout setup is valid or invalid regardless of where any single holder came in. The disposition effect is what happens when that private number starts to decide exits the plan had already defined. The way to find it is to compare what you wrote before the trade with what you wrote at the exit, using nothing the market showed afterwards.

Learn the pattern. Ride the trend. Keep the gains.

Educational content only. Not investment advice. Trading involves risk. You are responsible for your decisions.