Sunk Cost Fallacy in Trading: Why You Hold Losing Positions
The sunk cost fallacy is the pull to honor money and time already spent on a trade. See how it traps a futures trader in losing positions and how to catch it.
By Imperial Analytics
The sunk cost fallacy is the reason a trader adds to a losing position, holds it past the stop, and stays at the screen after a red day to win the money back. Each of those acts is built on money and time already spent, costs that the next trade cannot recover. This primer defines the fallacy, separates it from loss aversion, shows where it surfaces in a trading day, explains why the spent cost should not enter the decision, and gives a way to catch it in the trade log.
What the sunk cost fallacy is
The sunk cost fallacy is the tendency to keep committing to a position or a session because of money, time, or effort already spent on it. The cost is already gone and cannot be recovered by continuing, yet it pulls the trader toward throwing good capital after bad. Arkes and Blumer named the effect in 1985.1
A sunk cost is any cost that has already been incurred and cannot be retrieved by future action. In trading, the realized and unrealized loss on a position is sunk: it exists whether the trader holds, adds, or exits. The fallacy is letting that already-spent amount sway a decision that should depend only on what happens from here. The larger the prior commitment, the stronger the pull, which is the opposite of how a forward-looking decision should work.
The effect was first framed in economics by Thaler in 1980, who described how people treat money already paid as a reason to keep consuming what they bought rather than as a cost that is gone.2 The trading version is direct. A position that is down feels like an investment that deserves defending, when it is really just a current price and a forward question. The account does not remember what was paid to open the trade, and neither should the next decision.
How it differs from loss aversion
Loss aversion is the asymmetric sting of a loss as it happens. The sunk cost fallacy is the forward error that the sting produces: honoring what was already lost when deciding what to do next. Loss aversion explains the pain; the sunk cost fallacy explains why a trader keeps paying to avoid admitting it.
The two are linked but not the same. Loss aversion is about feeling: a loss weighs more heavily than an equal gain, so closing a losing position hurts more than the numbers alone suggest. That feeling is the fuel. The sunk cost fallacy is the decision that the feeling drives, the choice to keep the position open or to keep trading the day, specifically because so much has already been put in.
The distinction matters for the fix. Loss aversion cannot be willed away, because it is how the nervous system processes risk. The sunk cost fallacy can be blocked, because it lives in a decision, and decisions can be structured. A trader who accepts that the sting is permanent can still refuse to let the size of a past loss decide the next trade. That refusal is the entire skill.
How the fallacy shows up in a trading day
The sunk cost fallacy appears the moment a trade goes against the plan. It drives averaging down to lower the cost basis, holding past the stop because so much is already lost, and staying at the screen after a red day to win the money back. Each act is justified by what was already spent.
The clearest case is averaging down. A trade moves against the entry, and the trader adds contracts to pull the average price closer to the market, framing it as a better price rather than a larger bet on the same losing idea. The position is now bigger precisely because it was wrong, which is the fallacy in its purest form: the prior loss became the reason to risk more.
The second case is holding past the stop. The planned exit arrives, but the accumulated loss makes closing feel like locking in a defeat, so the stop is moved or ignored. The third case runs at the session level. After a losing day, the hours already spent and the money already down become an argument for one more trade to get back to flat, and that trade is taken outside any plan. The pattern is the same at every scale: a sunk cost is treated as a live reason to continue.
Data note
The averaging-down example in the next section uses illustrative round numbers chosen to show the arithmetic, not figures from any account or sample. Imperial Analytics surfaces a behavioral pattern from a trader's own data only when it meets the sample minimum in the AI Operating Charter, which is twenty trades in the matching condition.
Why the spent cost should not enter the decision
A trading decision is forward-looking: the only question is whether the next position has positive expectancy from here, at the current price. Money already lost is gone under every choice, so it cannot favor one choice over another. The correct test ignores the entry price and the running loss and asks whether the trade is worth taking now.
Consider an illustrative add. A trader is long two contracts of MES from a poor entry and is down -$300. The plan said to exit at the stop, which is here. The sunk cost framing says to add two more contracts to lower the average and recover faster. But the $300 is gone in both cases. The only real question is whether being long four contracts at the current price is a trade worth taking on its own merits. If the setup does not justify a fresh four-contract long right now, the add is not risk management, it is the fallacy doing the sizing.
This is why the entry price has no place in the exit logic. A position is worth holding only if a flat trader would open it again at the current price for the same forward reason. The running loss is identical whether the answer is yes or no, so it carries no information about what to do. Stripping the sunk cost out of the decision does not make the loss smaller. It stops the loss from recruiting more capital to defend it.
↳ Note
The account does not remember what you paid to get in. If a flat trader would not open the position here, the only reason you are still holding it is the money you already lost.
How to catch the fallacy in your trade log
The sunk cost fallacy leaves three marks in a trade log: positions added to while underwater, exits that land past the planned stop, and a cluster of trades late in a losing session. Tagging each entry as planned or improvised, and flagging any add to a loser, turns the pattern from a feeling into a count.
The log makes the fallacy measurable because each instance has a signature. An add to a losing position is a row with a second entry at a worse price than the first while the position was red. An exit past the stop is a realized loss larger than the planned stop distance. A make-it-back session is a run of trades after the day was already down, often increasing in size. None of these requires reading the trader's mind. They are arithmetic on entries, exits, and timestamps.
A simple practice surfaces all three. Tag every entry at the moment it is placed as planned or improvised, the same binary split covered in how to tag plan and improvised trades, and add a single flag for any entry that increases an already-losing position. Then review the improvised and added-to trades as their own group. If that group carries most of the day's red, the cost being defended is the sunk one, and the number is now on the page instead of in the story the trader tells after the close.
How to structure decisions so sunk cost cannot drive them
Three structural moves keep sunk cost out of the decision: set the stop and target before entry so the exit is decided once, ask of any open position whether you would enter it now at this price, and cap the session by a loss limit and a trade count so the screen closes before the make-it-back trade.
Pre-commitment handles the single trade. A stop and a target entered at the moment of fill convert the exit from a future emotional choice into a past mechanical one, made when the position carried no sunk cost yet. The decision happens once, before there is anything to defend, which is the only moment the trader is free of the fallacy.
The current-price test handles the open position. Asking whether a flat trader would enter here, for the same reason, replaces the entry price with the only price that matters. If the answer is no, the position is being held by its loss, not its setup. The session cap handles the day. A daily loss limit and a maximum trade count, set in advance, close the screen before the hours and the red already spent can argue for one more trade. The same evidence belongs in the review covered in what to review after a max loss day, where the added-to and improvised trades are the rows worth studying. Structured this way, the sunk cost never gets a vote, because every decision it would distort was already made before the cost existed.
Frequently asked questions
- q: Is the sunk cost fallacy the same as loss aversion? a: No. Loss aversion is the asymmetric feeling that a loss hurts more than an equal gain pleases. The sunk cost fallacy is the decision that feeling drives, continuing a position or a session because of what was already spent. Loss aversion is the cause; the sunk cost fallacy is the behavior to block.
- q: Why is averaging down an example of the sunk cost fallacy? a: Adding to a losing position to lower the average price uses the prior loss as the reason to risk more on the same idea. The correct test is whether the larger position is worth taking at the current price on its own merits. If it is not, the add is driven by the sunk cost, not by the setup.
- q: How do I tell a sunk cost decision from a normal scale-in? a: A planned scale-in is defined before entry, sized in advance, and triggered by the setup reaching a level, not by the position being underwater. A sunk cost add is improvised after the trade goes red and is justified by the loss already taken. Tagging entries as planned or improvised at the moment they are placed separates the two.
- q: Can experience remove the sunk cost fallacy? a: The pull does not disappear, but the behavior can be structured out. Setting exits before entry, asking whether you would open the position at the current price, and capping the session by a loss limit and trade count all remove the moments where the fallacy would otherwise decide. The skill is structural, not a matter of willpower.
Sources
Footnotes
-
Hal R. Arkes and Catherine Blumer, "The Psychology of Sunk Cost," Organizational Behavior and Human Decision Processes, Vol. 35, No. 1, 1985, pp. 124-140. ↩
-
Richard Thaler, "Toward a Positive Theory of Consumer Choice," Journal of Economic Behavior and Organization, Vol. 1, No. 1, 1980, pp. 39-60. ↩