Losing $100 feels worse than gaining $100 feels good, by a wide margin. That asymmetry is one of the most widely cited findings in behavioral science, and one of the most debated: the classic estimate is about 2.25, later meta-analyses land anywhere between roughly 1.3 and 2, and some researchers question whether a general asymmetry exists at all. Whatever the exact number, the direction of the effect is enough to produce the pattern every brokerage record shows: losers held, winners cut.
It is not a character flaw or a lack of discipline. It is how the brain is built, and the useful response is not to try harder but to build a process that does not depend on overriding it.
Where the finding comes from
Daniel Kahneman and Amos Tversky identified and formalized loss aversion, publishing Prospect Theory in 1979, a model of decision-making under risk that contradicted the standard economic assumption that people evaluate outcomes rationally against expected value.
Their central observation was that people do not evaluate final states of wealth. They evaluate changes relative to a reference point, and they weight losses and gains from that point differently.
The work reshaped economics and psychology, and Kahneman received the Nobel Prize in Economics in 2002. Tversky had died in 1996 and the prize is not awarded posthumously.
How large the asymmetry is
The classic coefficient, from Tversky and Kahneman's 1992 work, is about 2.25: a loss weighs a little more than twice as heavily as an equal gain. That figure is the one repeated in every textbook, and it is contested. Later meta-analyses disagree with each other. The largest, 607 estimates from 150 articles (Brown, Imai, Vieider and Camerer, 2024), puts the mean at about 1.96; a narrower one limited to fits on risky choices (Walasek, Mullett and Stewart, 2024) finds 1.31 and notes how few usable datasets exist. The estimate varies with the stakes, the framing and the population, and a minority of researchers doubt that a general asymmetry exists at all (Gal and Rucker, 2018; Yechiam and Zeif, 2025). The direction of the effect is robust; the size is not a constant.
The clean demonstration is a coin flip. Offered a bet that wins $100 on heads and loses $100 on tails, most people decline, despite the expected value being zero. Asked how much the win would have to be before they accept an equal chance of losing $100, the typical answer is around $200; the ratio measured in such experiments usually falls between 1.5 and 2.5.
One common explanation is evolutionary: for most of human history a loss of resources threatened survival while an equivalent gain merely improved comfort, so asymmetric caution may have been the right setting. It is simply the wrong setting for a market where outcomes are symmetric and probabilistic.
What it does to a trading account
Losers get held. This is the most expensive expression of it. Closing a losing position converts a paper loss into a realized one, and realizing is what hurts. So the position stays open, the stop gets moved, and the loss grows.
Traders describe this to themselves as patience or conviction. It is neither. It is a decision to avoid a specific feeling, purchased with capital.
Winners get cut. The mirror image. A profitable position creates the fear of giving the gain back (a loss relative to the new reference point), so it gets closed early to lock in the good feeling.
The combination is fatal. Small wins and large losses is a distribution that fails regardless of how often you are right. A strategy winning 70% of the time still loses money if the average loss is three times the average win, and loss aversion pushes the ratio in exactly that direction without anyone intending it.
Averaging down. Adding to a losing position to lower the average entry price. Occasionally a legitimate strategy, planned in advance. Usually it is an attempt to make a loss stop being a loss, and it converts a manageable position into an unmanageable one.
Revenge trading. After a loss, taking a larger or worse-considered position to recover it quickly. The reference point has become the balance before the loss, and everything is now evaluated against a number that no longer exists.
The disposition effect
The tendency to sell winners and hold losers has its own name in behavioral finance, the disposition effect, and it is among the most thoroughly documented phenomena in the field.
It has been measured directly in brokerage records across retail traders, professional fund managers and institutional desks, and across markets and decades. Investors are consistently more likely to sell a position that is up than one that is down, and the pattern survives the obvious alternative explanations: portfolio rebalancing, trading costs, tax-motivated selling.
The finding that makes it costly rather than merely interesting: the positions being held are, on average, the ones that continue to underperform. In Terrance Odean's 1998 study of 10,000 brokerage accounts, the winners investors sold went on to beat the losers they kept by about 3.4 percentage points over the following year. The disposition effect is not just psychologically uncomfortable; it is a measurable drag on returns.
The strange part: risk preference flips
Prospect Theory's most counterintuitive result is that people are not consistently risk-averse or risk-seeking. It depends entirely on which side of the reference point they are standing.
Facing gains, people become risk-averse. Offered a certain $500 or a coin flip for $1,000, most take the certain amount, even though the expected values are identical.
Facing losses, the same people become risk-seeking. Offered a certain loss of $500 or a coin flip between losing nothing and losing $1,000, most take the gamble, accepting more risk for the same expected outcome, purely to keep the chance of losing nothing.
This single reversal explains the pattern of a blown-up account. Up on the day, the trader takes profits quickly, because certainty is preferred. Down on the day, the same trader holds, adds, and takes larger risks, because the gamble is preferred to the certain loss. Risk-averse with profits, risk-seeking with losses, which is precisely backwards.
What actually works against it
The bias cannot be removed. It can be routed around, and every effective technique works by taking the decision out of the moment.
Hard stops, placed at entry. A stop loss decided before the position exists is decided by someone who is not currently losing money. Once the position is open and red, the person deciding is a different person with different preferences. Place it at entry, size the position to it, and do not move it away from price. The stop also has to exist in the platform: a stop entered with the order, at a price that caps the loss at, say, 1% of the account, is executed by the broker; a mental stop is executed by the person who is currently losing money, which is why mental stops move.
Predefine the exit on both sides. Target and stop, both before entry. This removes the profitable-position decision too, which is where the other half of the damage happens.
Think in probabilities across a series. A single trade has a binary outcome and invites emotional evaluation. A hundred trades has a distribution. Judging a decision by its outcome rather than its quality is the deepest version of this error, because a good decision with a bad outcome is still a good decision and should be repeated.
Size so that losses are survivable. Loss aversion scales with the size of the loss. A position risking 0.5% of an account produces a manageable feeling; one risking 10% produces the feeling that makes people move stops. Correct sizing is partly a psychological tool.
Keep a written record. Log the reason for entry, the planned exit, and what actually happened. Patterns become visible on paper that are invisible in memory, and memory is not neutral: it edits toward whatever protects the reference point.
Watch the reference point itself. The bias operates relative to whatever you have decided counts as zero. If that is your entry price, every position is evaluated against a number the market has never heard of. If it is your account balance at the start of the year, you will behave differently in December than in January for no market reason at all. The workable alternative is to make the current price and the written plan the only reference: for every open position, the question is whether it would be opened now, at this price, with no history, and if the answer is no, the history is not a reason to keep it.
The same bias outside markets
It is worth recognizing elsewhere, because the pattern is easier to see when money is not involved.
Marketing uses it constantly: limited-time offers exploit the fear of losing an opportunity, and free trials work because once you possess something, giving it up registers as a loss.
The endowment effect is a close relative: people demand more to give up an object than they would pay to acquire it, purely because they own it. The sunk cost fallacy is another: continuing a failing project because abandoning it would confirm the loss already incurred.
Holding a losing trade because you have already lost money on it is the sunk cost fallacy with a ticker attached.
Knowing the wiring
Loss aversion is not a weakness specific to bad traders. It shows up in professional fund managers, in experienced investors, and in the researchers who study it. Awareness reduces it slightly and does not remove it.
What separates traders who cope from those who do not is not willpower. It is that the coping ones have arranged their process so the decision is made in advance, by a version of themselves who is calm, and executed mechanically by a version who is not.
The stop was placed before the position existed. The target was set at the same time. The size was chosen so that neither outcome is emotionally significant. That is the whole technique, and it works because it never asks you to feel differently, only to have decided earlier.
