Role A
Play the Manager
You already paid the $100,000. You decide how to spend the next $50,000.
Week 3 · Practice Laboratory
A wasted cost. Two rational choices.
If money is already gone and unrecoverable, what does a self-interest benchmark predict — and what will people actually choose?
Private by design · no server, login, or tracking
The half-finished project
Your company has already spent $100,000 on Project X, and it is now clearly underperforming. Finishing it would cost another $50,000. The same $50,000 could instead launch Project Y, whose expected returns are much higher.
You must allocate the remaining 10 points (each = $5,000) between “keep funding X” and “switch to Y.”
Role B
Play the Advisor
Fresh eyes, no sunk cost. You see the same numbers and decide.
The self-interest benchmark
Because the $100,000 is gone either way, a strictly rational decider ignores it and puts all 10 points into Project Y — 0 into X. Sunk costs must not drive tomorrow’s choices. Emotional attachment or loss aversion may produce a different observation.
Step 2 · Decision
Where does the next $50,000 go?
Role A · Manager
Allocate the next $50,000
Drag the slider to split 10 points between continuing Project X and switching to Project Y.
Step 3 · Reveal
What the benchmark predicts
Rational benchmark
0 points → X · all 10 → Y
Sunk cost is ignored.
Your choice
0 → X · 10 → Y
Step 4 · Reflect
The sunk cost fallacy
- Only future costs and benefits matter. Money already spent is gone whether you continue or stop — it should never tip the scale.
- Decision-makers with sunk costs behave differently. The Manager (who paid) is systematically more likely to keep funding X than the Advisor (who didn’t), even with identical numbers.
- Good decisions are made from the outside. Ask: “If I hadn’t already spent the $100,000, would I still choose X?” If not, cut losses now.