Sunk Cost
A sunk cost is money, time, or effort that has already been spent and cannot be recovered, regardless of what decision is made going forward. In theory, rational decision-making should completely ignore sunk costs and weigh only the future costs and benefits of each option available now - but in practice, people (and traders) routinely let past spending influence decisions it has no logical bearing on, a well-documented behavioral bias known as the sunk cost fallacy.
A classic trading example: a trader buys a stock at $100, and it falls to $60. The original $100 purchase price is now a sunk cost - that money is already committed to the position one way or the other, whether the trader holds or sells. Continuing to hold purely because "I need it to get back to $100 to break even" is the sunk cost fallacy in action; a rational decision should instead ask, based only on current information, whether this stock is a good holding at today's $60 price, completely independent of what was originally paid for it.
Sunk cost is a psychology and decision-theory concept rather than a market-specific one, so it applies identically to forex, stocks, crypto, and commodities trading - anywhere a trader has already committed capital and faces an ongoing choice about whether to hold, add to, or exit a position. Recognizing the sunk cost fallacy is one of the more valuable pieces of trading psychology to internalize, since it directly targets one of the most common reasons traders hold losing positions far longer than their own analysis would otherwise justify.
This bias has been measured on real accounts rather than inferred from a laboratory. Terrance Odean examined 10,000 accounts at a US discount brokerage across 1987 to 1993, nearly 163,000 trades, and compared how often investors sold a holding showing a gain against how often they sold one showing a loss. Gains were realised at a rate of 14.8% and losses at 9.8%, meaning a winning position was more than half again as likely to be sold as a losing one. The pattern reversed in December alone, when tax considerations briefly made selling losers attractive.
The twist is what happened afterwards. The shares those investors sold went on to outperform the ones they held onto, by roughly three and a half percentage points over the following year. So the behaviour was not merely inconsistent with theory, it cost money, and it cost money in precisely the direction the theory predicts. The name for the pattern is the disposition effect, and it is the sunk cost fallacy wearing market clothes: the reluctance to sell at a loss is a reluctance to convert a decision that can still be rationalised into one that cannot. Noticing that the discomfort is about the conversion rather than about the money is most of the cure, since the money was already gone.