The Sunk Cost Fallacy: Why Traders Hold Losing Positions Too Long

In short. The sunk cost fallacy is the habit of letting money, time or effort you have already spent decide what you do next. In trading it shows up as holding a losing position because closing it would make the loss real. The loss is the same whether you close or not, so the only useful question is what the trade is worth from here.

What the sunk cost fallacy is

A sunk cost is something already spent that you cannot get back. Arkes and Blumer described the sunk cost effect in 1985 as a greater tendency to continue an endeavor once an investment in money, effort or time has been made (Arkes and Blumer, The Psychology of Sunk Cost). Because the past spending cannot be recovered, it should not change the decision, yet it often does.

Their theater experiment is a clean example. Season-ticket buyers were randomly given no discount, a $2 discount or a $7 discount. Over the first five plays, the full-price group attended more often, with a mean of 4.11 plays against 3.32 and 3.29 for the discounted groups. The effect had faded by the last five plays. The money was gone in every group, yet those who paid more kept showing up to justify it.

How it looks in a trade

A trade gives the fallacy plenty of room, because the "spent" amount is visible on the screen every second. Typical forms:

• Holding past the stop: the planned exit is hit, and the trader waits for price to come back to entry "to get out even".
• Moving the stop: the stop is pushed further away so the loss stays unrealized a little longer.
• Averaging down: adding to a losing position to lower the average entry, without that add being part of the original plan.
• Defending the analysis: hours of research or a public call make exiting feel like admitting a mistake.
• Waiting for break-even: the exit price becomes the entry price, not what the market now suggests.

None of these is a market signal. Each one is a reaction to the number already on the P&L line.

What the market data suggests

Odean studied trading records of 10,000 accounts at a large discount brokerage and found that investors strongly preferred realizing gains over losses. The pattern was not explained by rebalancing or by trading costs, and it was not justified by subsequent portfolio performance (Odean, Are Investors Reluctant to Realize Their Losses?).

Be careful with the label. That tendency is called the disposition effect, and researchers link it to several causes, including loss aversion, which is covered in how loss aversion distorts trading decisions. The sunk cost fallacy is a close relative with a different trigger: it is about what you have already put in, not only about how much a loss hurts. In practice the two usually travel together.

The forward-looking test

The fix is to ask a question that ignores the past. If you had no position today, would you open this one at the current price, with the stop you can actually place now? If the answer is no, you are holding for reasons that belong to the past.

Important. The loss is not the problem. Holding is fine when the original reason still applies and the risk fits your plan. The problem starts when the reason for staying is the amount already lost.

A worked example

The numbers below are illustrative, not real results. A trader buys 100 units at 100 with a planned stop at 97, a risk of 3 per unit. Price falls to 94. The planned stop was passed, and the open loss is 600, or 2R.

Option one: hold and hope. Option two: buy another 100 units at 94 to bring the average entry to 97, so price needs to recover only 3 points to reach break-even instead of 6. That looks safer, but the exposure has doubled. If price then falls to 88, the first 100 units lose 1,200 and the added 100 lose 600, a total of 1,800. Without the add, the loss would have been 1,200.

Averaging down did not change the market. It raised the stake on a trade that had already failed its own rule. If the trade was planned as a scale-in with a fixed total risk, that is a different case, and it is why the original plan matters.

Habits that cut the pull

• Set the exit before entry: write the stop and the reason in the journal before you click. A stop-loss order placed in advance takes the decision out of the moment.
• Size the risk first: if the stop loss is small relative to the account, it hurts less to honor it. See position sizing.
• Ask the fresh-eyes question: "Would I open this now?" Write the answer next to the trade.
• Never add to a loser by impulse: only a scale-in written into the original plan counts.
• Use a time stop: if the trade has not worked within the time you gave it, close it, even if it is not at the stop.

If the loss is followed by an urge to win it back, that is a separate trap, covered in how to stop revenge trading.

What to check in your journal

A journal turns this bias into something you can measure. Compare the average holding time of losing trades with winning ones: losers that last much longer than winners are a warning sign. Count how many trades closed beyond the planned stop, and how many had the stop moved. Add a short field, "reason to still hold", and read it in your review. A step-by-step routine is in how to review a losing trade.

On a prop firm account, a daily or maximum loss limit works as an outside stop that you cannot move. Holding a loser through it can end the account, as explained in what a daily loss limit is.

Common mistakes

• Calling every held loser a sunk cost: if the original thesis is intact and the risk is within plan, holding can be correct.
• Replacing the stop with a mental one: a mental stop is easier to cancel than an order.
• Averaging down without a total risk cap: each add feels small, and the total grows.
• Judging the trade by the P&L color: judge it by whether the plan was followed.

This article is for educational purposes only and is not financial or investment advice. The examples are illustrative and do not describe real trading results. Trading, especially with leverage, can lead to the loss of your entire deposit.

Log your reason to hold and your planned stop in the BitStat journal, and compare the holding time of winners and losers on the BitStat dashboard.