Loss Aversion and the Disposition Effect
Losses are felt more intensely than equivalent gains. The documented consequence is selling winners early and holding losers too long, which is the opposite of what the arithmetic wants.
MadStockAlerts Research · Updated August 28, 2026
What to take away
- A loss is experienced as roughly twice as intense as an equivalent gain.
- The disposition effect is the resulting tendency to realise gains and defer losses.
- It is well documented in real brokerage data, not merely in laboratory settings.
- A loss is not real only when it is realised; that framing is the mechanism.
- It attacks the arithmetic of expectancy from both ends at once.
- The disposition effect is the measurable consequence: winners are sold early and losers held, across large samples of real accounts.
MAD Academy Training Video · 0:45
Why Losers Get Held and Winners Get Sold
Losses feel roughly twice as strong as equivalent gains, which produces exactly the wrong holding behaviour in both directions.
This lesson is part of a Stock Alerts + Tools plan.
The finding
Prospect theory, from Kahneman and Tversky, found that people evaluate outcomes against a reference point and weight losses more heavily than equivalent gains, by a factor commonly estimated around two.
The consequence in markets is the disposition effect: a measurable tendency to sell positions that are up and to hold positions that are down. It has been documented repeatedly in actual brokerage records across many countries and account types.
It is worth stressing that this is a finding about people generally rather than about inexperienced people specifically. It appears in professional records too, in attenuated form.
Scroll the chart sideways to see all of it.
Why it is exactly backwards
Every method described in this library depends on losses staying small and gains being allowed to develop. The disposition effect does the reverse: it caps the gains and lets the losses run. It attacks the arithmetic of expectancy from both ends at once.
The recovery table in the risk pillar is the arithmetic being attacked. A loss allowed to deepen from 10 percent to 50 percent moves from needing an 11 percent gain to recover to needing 100 percent, and every day the position is held is a day the bias is winning.
The framing that sustains it
The thought that a loss is not real until it is realised is the specific belief that keeps a losing position open. It is also false in every sense that matters: the capital has already declined, and the account statement says so.
Whether the position is held or closed changes nothing about that, and only changes whether the capital is available for something else. The unrealised loss is a decision to keep making the same bet, restated as a decision not to decide.
The conventional countermeasures
- Deciding the exit before entry, when the position is not yet a loss and the decision costs nothing.
- Using a resting order, so the exit does not require a decision at the moment it is hardest.
- Asking whether the position would be entered today at this price, which reframes holding as a fresh decision.
- Recording the reason for each entry, so the reason's disappearance is noticeable.
The third is the most portable, because it needs no infrastructure. Holding is a decision to buy at the current price, and phrasing it that way removes the reference point that the bias depends on.
The disposition effect, which is the measurable version
Loss aversion is a description of how outcomes feel. The disposition effect is what it does to behaviour, and unlike the feeling it can be counted: across large samples of brokerage accounts, holders sell winning positions markedly sooner than losing ones, and the effect is one of the most reliably reproduced findings in the study of individual investors.
The mechanism follows from the shape of the value curve. Realising a gain converts an unrealised win into a certain one, which is pleasant; realising a loss converts a paper loss into a permanent one, which is the moment the loss becomes real. Holding therefore feels like preserving optionality, when arithmetically it is only continuing to hold a position that has moved against the reason for holding it.
The compounding problem is that this is exactly backwards relative to how prices behave over most horizons. Positions are cut short precisely where they are working and held precisely where they are not, and the aggregate effect is a distribution of small wins and large losses regardless of how good the selection was.
The finding is about averages across many accounts, and it says nothing about any particular decision. What it does establish is the direction of the pull, and the direction is consistent enough that a process which does not account for it is relying on being unusual.
Why the entry price keeps mattering when it should not
The price paid is a fact about the past and it carries no information about what the security is worth now. Every participant with a different entry price faces exactly the same set of future outcomes. Nevertheless the entry price is the single reference against which most holders evaluate a position, and it is the reference the phrase break even points at.
The reason is structural: the entry price is where the felt curve crosses zero, so it is where the asymmetry between gain and loss is anchored. Everything above it is measured on the shallow half of the curve and everything below on the steep half, which is why moving from minus eight percent to minus two feels like a larger event than moving from plus two to plus eight.
- Waiting to get back to break even is a decision to hold based on a number no other participant can see.
- A position that has fallen is not owed a recovery by anything, and the shares do not know what was paid for them.
- The question the position poses is whether it would be opened today at today's price, which is a question the entry price cannot answer.
- Averaging down changes the reference rather than the position, which is why it feels like an improvement even when the exposure has grown.