Sunk Cost Bias
Learn how sunk cost bias keeps traders committed to losing positions and research, and how forward expected value and opportunity cost improve capital decisions.
Reading progress — saved on this device
Sunk cost bias is the tendency to keep committing capital, time or attention to a trade because resources have already been spent, even when the forward expected value no longer justifies the commitment.
Learning objectives
- Distinguish sunk costs from relevant forward costs and benefits.
- Evaluate adds and exits from current conditions.
- Use thesis reviews and capital-allocation comparisons to reduce commitment escalation.
What it is
Money already lost on a position cannot be recovered by changing the past. The next decision is whether the current position, at the current price and with current information, deserves the capital it occupies.
Research effort is also sunk. A trader may spend weeks studying a project and then feel compelled to own it so the work was not wasted. Research can still be valuable when the correct investment decision is no position.
Sunk cost bias often combines with anchoring and loss aversion. The trader focuses on getting back to entry rather than comparing the position with alternative uses of capital.
How sunk costs distort capital allocation
Use an inheritance test: if this exact position appeared in the account today at the current price and size, would you choose to keep it? This removes the story of how the position got there.
When averaging down, calculate the new marginal trade separately. Lowering average cost is an accounting effect. The additional capital should still have a valid expected payoff and fit the portfolio risk budget.
Opportunity cost makes the bias measurable. Capital tied in a broken thesis cannot be allocated to stronger setups or held as cash or stable reserves.
Predefined thesis-review dates or invalidation events reduce the tendency to move the goalposts after losses. The review should ask what evidence changed, not how painful realising the loss would feel.
Control framework
| Check | Purpose | What to verify |
|---|---|---|
| Forward thesis | Removes past commitment | Restate expected return and risk using only current evidence. |
| Inheritance test | Reframes ownership | Ask whether you would buy or hold this exact position today. |
| Marginal add | Tests averaging down | Evaluate new capital independently of the existing loss. |
| Opportunity cost | Compares alternatives | Rank the position against cash and other qualified setups. |
Worked example and thought exercise
A trader invests 10,000 in a token that falls 50 percent to 5,000. They consider adding another 5,000 because they are already "in too deep." The previous loss is sunk. The relevant question is whether a fresh 5,000 invested at today's price has better expected value than other alternatives.
If the project's core adoption metric has collapsed and the original catalyst was cancelled, adding simply to reduce average entry can increase exposure to a weaker thesis.
Thought exercise: why can closing a losing trade be a rational decision even if the asset later rebounds?
Common mistakes and practical workflow
- Holding because "I have already lost too much to sell."
- Adding only to lower average entry price.
- Treating research time as a reason to maintain exposure.
- Refusing to compare a legacy position with new opportunities.
Practical workflow
- Ignore historical P&L temporarily and restate the current thesis.
- Apply the inheritance test at the current size.
- Evaluate any additional capital as a standalone decision.
- Compare expected payoff with alternative uses of capital.
- Exit or resize when the current thesis no longer justifies the opportunity cost.
Knowledge checkpoint
- What makes a cost sunk?
- Why is lower average entry not proof that an add is attractive?
- How does opportunity cost expose sunk cost bias?
- What is the inheritance test?
FAQs
❓ Is a realised loss economically worse than an unrealised loss?
Current market value already reflects the loss. Realisation can matter for tax and accounting, but not because the market owes a return to entry.
❓ Is averaging down always wrong?
No. It can be rational if the updated thesis and risk justify the new marginal purchase.
❓ Why is research time a sunk cost?
Because time already spent cannot be recovered by owning the asset.
❓ What is the inheritance test?
Ask whether you would choose to own the same position today if it appeared in the account without its history.
Summary
Sunk cost bias is reduced by evaluating every position from today forward. Prior losses and prior effort are history; current expected value and opportunity cost determine whether capital should remain committed.
Want this in a personalised order?
Take the crypto assessment and get a custom path of 10 modules matched to what you already know. Free, no card required.
Build my path →