Is the so called sunk cost fallacy truly a total fallacy, or does it have some kernel of truth? Certainly it is a widespread instinctive/impulsive form of reasoning. Presumably it must then have had some evolutionary advantage that its application bestowed on organisms who would embrace it. What is this?

  • Truly a fallacy?
    – Boba Fit
    Apr 14 at 20:24
  • 6
    If we go by the dictionary description, "the phenomenon whereby a person is reluctant to abandon a strategy or course of action because they have invested heavily in it, even when it is clear that abandonment would be more beneficial", then yes, it is a "total" fallacy (or rather folly). The problem is that in realistic conditions it is not always "clear" whether abandonment is more beneficial or not. In conditions of future uncertainty, persistence may well have evolutionary advantages, but psychological habits tend to spill over even into situations where they are clearly irrational.
    – Conifold
    Apr 14 at 21:58
  • 2
    The idea that any and all characteristic that perdurates through the ages must be beneficial in some way is a gross misconception of evolution. To perdurate a traits does not have to be beneficial, just be not detrimental enough that individuals who have it can live long enough to reproduce. We all have a blind spot because our optic nerve connects to our retina on the wrong face, it has no benefits, and some animals don't have it, sign that it is not some kind of physiological necessity. But it's not enough to kill us before we have kids, so it remains.
    – armand
    Apr 17 at 9:17

6 Answers 6


I would say it's not a total fallacy. If you're 90% done with something meaningful but you're suffering, quitting might be or might not be a good idea, depending on the activity.

  • “Might be or might not be “ … can you elaborate or expand on that at all? Apr 14 at 20:39
  • 1
    If you're watching a movie and you're sure you don't like it at all and you're 90% done, it's a good idea to quit. But if you're 90% done with a college degree and you're suffering, it's a good idea to just push through it and not think of "sunk cost fallacy". Apr 14 at 20:41
  • Well, if it is better to continue then it’s not a fallacy. If you spent 90% of the money needed, but the value of the result is less than the missing 10%, then continuing is a fallacy.
    – gnasher729
    Nov 28 at 17:39
  • People keep arguing scenarios where it is basically a "total fail", that there would not be any reasonable return on the entire investment, not just that the last bit wouldn't pay off. These are scenarios where the project never would have been considered if more information was available earlier. Most sunk cost situations are not that extreme, but unfortunately some are. If you don't like the movie you are watching, maybe decide in the first 10% instead of the last? On the other hand, I got two thirds through a famous book I was enjoying when I hit something that caused me to quit it.
    – Scott Rowe
    Nov 29 at 11:50

According to this article (thedecisionlab.com), the sunk cost fallacy stems from the commitment bias. How so?

In the context of cognitive science, a fallacy might be viewed as a defect in reasoning that is prompted by a cognitive bias. For instance, an argument that appeals to popularity might expose a bias to relying on testimony as an epistemological source. Robert Audi, in fact, in his introductory text on epistemology cites five sources: consciousness, perception, memory, reason, and testimony. If we have a psychological susceptibility to conforming to group think, as Solomon Asch's famous experiment tends to confirm, then the susceptibility to selecting testimony over reason is very much in conformance with those in evolutionary psychology and behavioral economics (such as Tversky and Amos) who argue that our biases offer some sort of tradeoff; clearly, in hunter-gatherer societies, where in-group-out-group thinking dominates, one might be able to show that reason is far less important than social conformity. The book On Bullshit would count as evidence towards, that of rhetorical forces, logos may be the least persuasive relative to logos and pathos, if you buy rhetorical theory going back to the Ancient Greeks. (Watch Antz for a playful expose on sociobiological thinking.)

What are the philosophical implications? Well, that psychologism is better than anti-psychologism, and that logic is a subset of the psychological which has certain implications in philosophy of mind, where the question of the metaphysical necessity of a great many things is argued, including what metaphysical necessity itself is. The one caveat that should be published is that this response, this line of reason in philosophy presupposes a naturalized epistemology (SEP) that is heavily indebted to cognitive science. There are philosophical opinions and entire communities who approach the notion of truth and fallacy from a less robust scientific perspective and might rely on more historical philosophical notions.


It is perhaps worth starting with the question "What is a fallacy?" A fallacy is a connection in a logical argument which does not necessarily follow. There is in fact a wonderfully named "Fallacy fallacy" whereby someone assumes the presence of a fallacy means the conclusion is necessarily false. Most of the informal fallacies are exactly as you allude to; typically helpful rules of thumb. Without making any claim about any genetic basis for thought patterns, you are correct that they are patterns which could "evolve" at least socially. For example, someone who in making a snap decision, appeals to the wisdom of experts or of consensus is likely to do better than someone who doesn't, even though it is logically possible that experts or common understanding may be wrong.

Now, what is the sunk cost fallacy in particular? The Sunk cost fallacy happens when new information reveals that some course of action is not worthwhile, and the person sticks at it anyway. Here are a few reasons that such a pattern may, as you put it, evolve:

  • Prevarication is costly. That is, continually reevaluating decisions can be so expensive that it can be cheaper to settle on a decision already made. For example, a business with a few employees working on a dead end project may be more efficient than one where all employees are constantly in meetings asking "Are we sure?"
  • Progress is usually valuable. To be an example of the fallacy, the expected reward must become lower than the expected remaining cost. Let's put some numbers on this for an example: A company starts on a 100k project which they reasonably expect to bring in 300k revenue. 80% of the way through they encounter new information which suggests it will only bring in 50k revenue. It's tempting to say that they "Shouldn't have started" because that revenue is less than their 100k cost (although in fact they made the right decision with the information they had at the time). However, it's probably still the right decision to finish the project and spend the remaining 20k for the 50k return. It's a fallacy to count that they have "80k worth of progress" since the amount originally paid is irrelevant. It is not a mistake to count that they have "80% of progress" and that can swing the balance. Of course if the numbers were different then abandoning it may be the better choice, but it's easy to see how subtle these things can be.
  • The decision maker may not bear the full cost. In particular, people are notoriously bad at distinguishing "A reasonable bet that didn't work out" and "A bad idea." Because of this, the social cost to a decision maker folding is higher than it should be. While a business may prefer to scrap a project when it becomes clear that it only has a 30% chance of success in terms of expected profit, it may be better for the manager of that project (considering her reputation and personal chance of promotion or dismissal) to hold course.
  • Almost everything worth doing has some moment where it seems intractable. Perhaps it's the frustration of getting stuck on a critical step in a maths puzzle or the pain of a cramp in the final stretch of a race. A certain amount of raw pigheadedness can be the only thing which gets people through these very visceral obstacles. Perhaps a perfect robotic logician would evaluate differently, but in situations where logic shuts down because "It just hurts" it can be helpful to have a non-analytic faculty capable of going "I don't care. I'm doing this."
  • Right, I am a successful programmer mainly because I didn't give up many many times, almost to the point of pigheadedness. There's no benefit to quitting.
    – Scott Rowe
    Nov 29 at 1:07

There are situations where you have the choice: Either admit that you made a costly mistake or spend more money on the mistake.

If admitting a mistake is a cost to my reputation and the cost of continuing is paid by someone else (let’s say my company, or a customer, or the tax payer), then to me it’s a good strategy to continue.

  • This pretty much explains the USA.
    – Scott Rowe
    Nov 29 at 11:57

From a scientific stance, I think it is a form of "good enough" reasoning, which confers an evolutionary advantage. From a philosophical point of view, I do not consider it a fallacy. There will be a probability that the assumption that the investment cost justifies further investment is correct. Whether any particular decision at any particular time is folly is contingent.

  • 1
    There is an old saying: "Better not to begin, but once begun, better to finish." I think it was about pursuing self-realization though. I would say it is always good to begin and persist in that.
    – Scott Rowe
    Nov 29 at 11:55

Is the so called sunk cost fallacy truly a total fallacy, or does it have some kernel of truth? Certainly it is a widespread instinctive/impulsive form of reasoning.

Gambling addiction is an excellent example of this fallacy and the outcome dependency.

A gambler keeps pouring money into a slot machine even though the odds are slim because he has already invested and the machine is "ready to payout"

There are two outcomes:

  • Jackpot - No fallacy.
  • Out of money, pawned everything - Fallacy.

The sunk cost fallacy is psychological as well which may explain why the behavior persists. Additionally, the belief that winning a Jackpot somehow negates the fallacy helps perpetuate the behavior.

In product development its called "feature creep": the idea that adding new functionality to a losing product will help revive it rather than designing a new product.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .