Arkes and Blumer, 1985
Money and time you cannot get back bend the next decision. Yes, sitting through the bad film.
The classic experiment: you bought two ski trips and the dates clash. One cost 10,000 yen, the other 5,000. You know the cheaper trip is the better trip.
Most people go on the expensive one. Because it cost more.
Which is strange, since the 15,000 yen is gone either way. The only usable input is which trip is better, and the amount already paid wins instead.
Three forces. You do not want to book the loss — quitting is the moment "wasted" becomes final, and while you continue you have not lost yet. You want to be consistent; reversing your own decision means admitting an error. You are being watched, and nobody wants to look like a quitter.
None of it is rational and all of it is human, which is why intelligence does not protect you. Only procedure does.
Ask it fresh: "if this landed on me today, would I start?" If not, there is no reason to continue.
And set the exit before you begin: leave if it is not good after thirty minutes; stop if there is nothing in three months. Decided in advance, quitting is a plan rather than a failure.
Same idea as a stop-loss — and nobody sets one successfully once it already hurts.
The last plate at an all-you-can-eat
You are full and you keep eating to get your money back. The money is gone, so the nausea is pure additional loss.
The useful thing about this example is that the loss is felt physically.
The game you spent money on
You keep playing an unfun game because quitting seems to waste what you spent. It does not come back if you continue either. You spend more time protecting money that is already gone.
The job or the relationship
Where "after all this time" hits hardest — and years tell you nothing about what happens next.
But calling everything a sunk cost is its own crudeness. If what you built still produces value, it has not sunk. The question is only what it produces from here.
If you take one thing
The moment you feel "but I already paid" is the moment the judgment starts bending. Stop there and recount using only what is ahead.
Related