Sunk Cost Fallacy

Villager: A lot of people gave very selflessly to build this warship so we can go out and battle the Vikings. But the time has come to admit that hard work and hope are no substitute for actual knowledge. And that we've made a really shitty ship. If we sail this ship against the Vikings, we'll be massacred immediately. I suggest we break it up for scrap and never speak of it again.
Other Villager: Throw away months of work? Fuck that! Let's fight! [crowd cheers]
[cue burning ship]
Vikings: [incredulous] Seriously, how does a boat just catch fire by itself?
Oglaf, "Bilge"

Sunk cost fallacy is a cognitive bias that causes Bob to remain committed to a course of action because he's already spent time or resources on it, even though the commitment is irrational (i.e. he would be better off walking away). When Bob is engaging in this fallacy, he will remain set on the course of action even if the profit from his success would be less than what he's already spent.

His line of thought might run thus:

I have already spent eight dollars on this contest. If I spend seven more, I can win the prize.
I could buy the prize elsewhere for five dollars.
But if I stop now, the money I already spent will be wasted, so I must continue.

This trope is not merely Bob's commitment to a course of action because he's invested too much to turn back. If Bob has already made the down payment on a house, for example, he is likely to continue in the purchase of that house since he would lose money for no gain if he stops. In order for a situation to be the Sunk Cost Fallacy, it must be one where (1) Bob remains committed to a course of action in order to justify what he has already spent on it and (2) it is obvious to any rational person that the cost of staying in now will exceed the cost of simply stopping now, taking on the loss, and moving on.

Compare Know When to Fold 'Em, I've Come Too Far.

Also called

  • Foot in the door (in sales)
  • Camel's nose (in social psychology)
  • Throwing good money after bad (in business)
  • Pay-or-play (in sports and entertainment)
  • Pouring sand down a rat hole (usually in politics)


  • In psychology this is generally related to cognitive dissonance theory, which is basically the idea that when people think of themselves one way but act in another way, they will try to rationalize it. Specifically, people think they make good decisions, but when the decisions aren't paying off, they throw more effort after it to make the decisions good rather than admit they made a bad decision.
  • It's also known in social psychology as a great way to bring someone into a group. Cults are known for using this: How about you read a flyer? Sure, that cost nothing. Hey, why don't you answer this quiz on how happy you are with your life? Well... You've already read the flyer, that's not much more effort. How about going to a session? Once the new recruit has spend hundreds, if not thousands of dollars on seminars and training, they are unlikely to be willing to go cold turkey; cognitive dissonance becomes involved (as quitting would require admitting to themselves that they were duped, which is something no one wants to think about themselves).
    • Church of Happyology is infamous for this,what with having to fork over $100,000 in the first fucking year.
  • More serious (and much more complicated) in life or death matters. Dead troops are a sunk cost, but many nations across history have continued on losing military campaigns because their leadership could not face up to having wasted troops.
  • This is often the reason people will spend lots of time on internet arguments, even after they're losing. Or as Scott Adams put it, "Nothing makes [someone] argue harder than being proven wrong."
  • This is one of the main reasons why people continue to play MMOs even when they aren't enjoying them.
    • "Free-to-play" MMORPG games. After sinking substantial money into the game (to get powerups that nonpayers wouldn't have), the player feels compelled to continue, even when the grind is getting difficult. This is especially true of games where experience (or worse, levels) are lost for death. You pay up, or one lag-related death takes an enormous amount of experience. Despite getting to the point where even 0.10% experience takes hours while death loss takes only seconds, the player continues to play. It's worse in some ways than for subscription-based, since supposedly "nobody forced you to pay" (although the game itself may be balance-weighted toward payment, by making items scarce and leveling slow), making you feel personally committed rather than having paid what was effectively a usage cost.
    • Similarly, people who don't enjoy playing a particular video game will keep playing it because of all the time they invested in it in the first place, even if the game is full of bugs, has bad storywriting, etc. This is usually along the lines of "Well if I don't beat the game, all the time I invested in it would be a waste." This is also the similar argument from people who try to beat the game on the hardest difficulty and refuse to quit no matter how many times they lose or how many hours they spend repeating the same sections in the game over and over again. Likewise, people who complain about every change that a developer makes in their game via patches tend to keep playing, despite being so vocal on why they hate the game instead of looking for another game that would suit their needs.
  • In poker, some players will continue to play poor cards just because they've already sunk a lot of money into the pot, even when there's no realistic chance that they'll win. Technically speaking, being "pot-committed" is not this fallacy; see below under "looks like this fallacy but isn't."
  • One example of this is the dollar auction. An emcee decides to auction off a dollar with a starting bid of one cent (which may be adjusted for inflation) - but there's a catch. The high bidder gets the dollar, but the second-highest bidder still has to pay their bid and gets nothing. The bidding will start off with each of the bidders standing to profit, but once the high bid reaches 99 cents, the second bidder has to choose between losing 98 cents or bidding one dollar and making nothing. After this, the first bidder has to choose between losing 99 cents or bidding $1.01 and losing a cent. This process of bidding will continue even though neither side stands to gain from future bids, except that by 'winning' they lose about one dollar less than if they had lost.
    • Interesting game. The only winning move is not to play.
    • Auction websites such as Quibid and Beezid work in a similar fashion, except that you have to purchase your bids for a much higher amount than the bid increments. To bid a price up by 10 cents, you might have to spend five dollars.
  • Happens in labour disputes where management or unions try to recoup the losses from a strike or lockout, and that merely pushes the bargaining positions of the parties further apart.
  • Within sociology, the combination of this and the Madonna–Whore Complex often leads to girls feeling resigned to becoming sexually promiscuous after having had sex once. After all, they can only lose their virginity once and what does it matter after that? Good sex ed would help prevent that, as the risk of catching an STD or having to deal with unwanted pregnancy increases with each sexual encounter.
  • The Dutch public transport card system started out as a bad idea, turned out to be a worse idea, became a giant money sink, and still the government refused to just fess up and admit their mistake. A few years later, it's currently being debated as the worst thing that ever happened in the history of Dutch transport. Attempts by the government to just buy out the company responsible for the mess are still failing miserably.
  • A common example is an automobile owner near the end of their vehicle's functional life. So many of us end up trapped in this Fallacy and wind up spending good money on repairs to stretch out the life of a car that's already on its last legs, because to not keep driving the thing after investing so much in it just seems like a waste.
  • Stephen Colbert, on The Colbert Report, summed it up quite succinctly when discussing the American dilemma of whether torture was justified since it (allegedly) helped to capture and kill a hated terrorist. Stephen's usual Insane Troll Logic is applied to the point where, because America has already lost its beloved moral superiority by using torture, they have to keep torturing until it solves all of our problems, or:
  • Supporters of The Vietnam War and the U.S. occupation of Iraq said that unless the U.S. continued the wars, the lives of soldiers who'd already died there would be wasted. Notably, during the troop surge for the Iraq War (a move that was deeply unpopular when it was first proposed), Bush deliberately used this exact phraseology. Lo and behold, support for the move suddenly shot up dramatically.
    • The World War One poem In Flanders Fields by John McCrae echoes this sentiment, with the ghosts of the fallen telling the living to keep fighting so they will not have died in vain.
    ''Take up our quarrel with the foe:
    To you from failing hands we throw
    The torch; be yours to hold it high.
    If ye break faith with us who die
    We shall not sleep, though poppies grow
    In Flanders fields.''
  • 419-scams work this way. A good scambaiter is able to turn the tables on the scammer, usually with hilarious effect. (Imagine the scammer thinking: "I've already posed for a photograph of me with a loaf of bread on my head, carved a wooden replica of a Commodore 64 keyboard and gotten a tattoo reading "Baited by Shiver". Surely this priest will give me $270000 soon!")
  • Gambling addictions usually fall into this trope. While people who are hooked on gambling are hooked on the rush of risking everything to win big, other addicted gamblers will gladly keep blowing money on a game until they can win back everything they lost just because they already lost money in the first place. For example, if someone were to lose $5000 in a game, they will keep spending money on that game until they can win back that $5000 plus the additional money spent to get back the initial losses. In other words, "I already invested so much money in this game, I may as well keep playing until I can win everything back." Overlaps with Gambler's Fallacy, because obviously a losing streak means you have to win something soon, maybe the very next game.
  • This phenomenon is also referred to as the "Concorde fallacy". The Concorde supersonic airliner was enormously expensive to develop and maintain, but never truly turned a profit in service. It was only kept in service by Air France and British Airways because to do otherwise after investing so much in the development of the plane would have been wasteful.
  • Happens in sports a lot. When a big-name player signs a big-money contract then starts sucking, a lot of teams will continue to play him rather than exploring other options. Granted, this might have something to do with perception (signing a guy to a huge contract then benching him immediately would look very bad).
    • There are also merchandising aspects to consider (perhaps playing him will leading to increased sales items with the player's name/likeness at the gift shop of).
    • They are likely going to have to paid either way and count against the salary cap as well (if there's one in place).
    • There is usually a high chance the team is operating without a good back-up player (or the back-up only possesses a functional knowledge of the team's playbook) if they're willing to invest so much of their time and money in their star player.
    • There is also the concept of a Replacement Player, or the guy you can get in uniform right away. Sports have a long-tailed, skewed distribution of talent. The key idea is that most professional players are below average, and so average and better players are hard to get. Replacement players, however, are readily available. A former superstar having a mediocre season is still usually better than average - let alone whoever can be readily found to take their place. So it can still make sense to keep that fallen superstar on staff.
      • Alex Rodriguez, the poster-boy for overpaid, past-their-prime star, is a good case study. As of early June 2015, he is putting up an On-Base Plus Slugging (OPS) of 0.880 (a rough measure of offensive contribution). The American League average in 2014 was 0.706, making A-Rod still better than average. This would rate him a Very Good hitter in most rating schemes. A-Rod has never had a season where he has been average or worse as measured by OPS. More advanced metrics exist, and they generally show that A-Rod was always worth a few wins above an Average player (except in 2013) and worth a lot more than a Replacement Player. In that case, keeping A-Rod's overpriced bat in the line-up is not this fallacy if you're stuck with him - or if you happen to have a massive cash reserve and are willing to blow it for marginal gains.
  • Any large-scale project one party already invested large sums of money in only to see the project needing more and more cash due to unforeseen problems, missed deadlines, etc.
  • How many people (probably even some reading this very page) have been (or currently are) in a relationship that they know is unhappy, unhealthy, and/or isn't going anywhere, but decide to stay because they've known their partner for so long and can't imagine life without them? This especially applies to abusive relationships, as the abuse victim has poured a lot of energy into trying to change the abuser, an effort they think will be wasted if they leave.
  • Competitions whose prize money is based around a pot can easily lead players into this. By paying some certain amount, a player is allowed to compete for a bit. If there are competitors already ahead of him or her when time runs out, he or she is likely to pay more to continue to compete in hopes of beating those people. If he or she is already on top, the player is still more likely to pay more out of fear that the competitors could outdo him or her when he or she is not playing. The competitions set up by the Professional and Amateur Pinball association function like this, and competitive players often sink hundreds of dollars in hopes of getting that higher score on that one machine. (Such competitions have a strong overlap with Gambler's Fallacy though, due to the high luck factor pinball has.)
  • Aresh, the Arc Villain of Jack's loyalty mission in Mass Effect 2, uses a version of this to explain why he wants to restart the Pragia facility - the people running it did such horrible things to everyone there, including him, that they must have been pursuing some greater purpose, and its near-destruction during a riot means that all the pain that happened there was for nothing. Jack, meanwhile, is understandably scornful of the idea that the suffering caused at Pragia meant anything more profound than "the people running this place need to die".
  • In Conned Again, Watson! by Colin Bruce, a book that uses the Sherlock Holmes characters to explain statistics, Holmes explains this fallacy and several others to the client, a businessman who is putting more money into a less successful venture because it cost more to start with, and Watson thinks smugly that he can avoid mistakes like that. The following day, Watson buys a loaf of bread for sandwiches, and then walks past a street vendor selling ham sandwiches at a reduced price to get ham from the butcher. When Holmes points this out, Watson, of course, replies "But I had already bought the bread."
  • When done badly, the policy doctrine of 'too big to fail' can become this. One such example was the car maker British Leyland, which was bailed out by the British Government in 1968, but still failed to reverse its long-term decline. Another example is the general concept of corporate welfare.
  • The Order of the Stick's Redcloak suffers from this, as expressed in Start of Darkness; It's not that he believes in the Plan as much he believes that if he gives it up, it'll make all of the horrible things he's done worthless, in spite of being told from both his brother and Xykon himself what an empty excuse this is. Also, he continues to support Xykon despite being entirely too familiar with the lich's Bad Boss habits and knowing that completing the Plan with Xykon will not work out in his favor, he feels too invested to quite and find some other, saner arcane spellcaster to work with instead.
  • Similarly, Locus from Red vs. Blue. He has to keep following orders because admitting he has a choice would mean admitting he had a choice in all his previous actions.
  • Reacher Gilt's scam in Going Postal relies heavily on this. Even as the service on the Grand Trunk semaphore line gets worse and worse, he sweet-talks investors and board-members into "throwing good money after bad" while pocketing most of it and covering it up with tricky accounting and corporate buzzwords like "embracing diversity" and "synergistically". Not-quite-reformed con artist Moist von Lipwig is equal parts impressed and disgusted when he realizes what Gilt is up to.
  • In the Triptych Continuum, this is the central reason the Crusade continues: the Cutie Mark Crusaders have essentially become fanatics in the name of their cause, and to stop would invalidate everything which had come before — including all the accidents, disaster relief forms, reparations, and tree sap. The Fallacy itself winds up being summarized in a speech by Apple Bloom, and the voicing of it is part of what finally breaks the Crusade.
    "Gotta do it for a day because y'try," she slowly said. "Then y'go for a week 'cause iffin y'don't, y'wasted the day. Then it's a moon, lots of moons, and then when it's a year, it's gotta be more. We keep goin' an' goin' 'cause if we ever stop, then it means we wasted everything. An' we could just keep goin' til we're grown up, out of school, but we won't have jobs because we don't have marks and the only thing we can do is look some more. It's nearly three years an' if Ah do it for one more day, it could turn into... It's too much, an' Ah think — it's been too much for a while. Too long. An' — an' it ain't worth it no more."
  • At the end of a "No Mercy" run (i.e. a "kill everyone in the game including and especially major characters" run) of Undertale, the Final Boss points out that you're pressing forward with your murdering spree not out of any sort of "good" or "evil" desire, but simply because you can, and because you can, you feel like you have to, even though there's no real benefit to persevering at this point. They also gladly point out that the best thing to do at this point is to seriously give up and do literally anything else. Because really, what do you personally have to gain from not only massacring the entire underground but then also destroying the entire world and selling your soul to reset the game at the cost of ruining all future good endings?
    i always thought the anomaly was doing this cause they were unhappy. and when they got what they wanted, they would stop all this. and maybe all they needed was... i dunno. some good food, some bad laughs, some nice friends. but that's ridiculous, right? yeah, you're the type of person who won't EVER be happy. you'll keep consuming timelines over and over, until... well. hey. take it from me, kid. someday... you gotta learn when to QUIT. and that day's TODAY.

Looks like this fallacy but is not:

  • Any situation in which the choice to go with option A over option B is based purely on how much getting to the desired result through either option is expected to cost from now on, regardless of how much money, time or other resources have already been spent on option A.
    • Remember, falling for the reverse is just as bad. If option B only costs $5, and option A costs $15, of which $14 has already been spent, option A is the better choice, no matter how cheated you feel over the difference. The choice here is still made based on how much both options will cost from where you are right now.
  • When abandoning the current plan has costs that outweigh the benefit of switching to a better plan; for example, a penalty clause for cancellation of a contract that is higher than simply paying the contracted price until the contract runs out. Cell phones and cable/satellite services, health clubs, and auto leases often have these. For example, a cell phone contract is 2 years at $20.00 a month, and has a $250.00 cancellation penalty. If 12 months or fewer remain on the contract, it costs more to cancel than it does to simply continue paying the contracted amount until the contract expires. Another example would be, if in the contest above, the person had spent $11 rather than $8. Assuming victory was certain at $15, continuing to play would be a reasonable decision. Continuing to play costs $4 more, making $15 total. Stopping after spending $11 and simply buying the prize elsewhere for $5 costs $16 total — so why stop?
    • Or if the contest itself is something fun enough to be worth at least $2 in its own right. This is how things like carnival prizes work.
    • Or the contest is for charity, so even if you lose the money you spend is going to a good cause.
    • The above examples simply point out that the Sunk Cost fallacy should account utility, not merely monetary value. Utility is an abstract and impossible to objectively measure concept, but it is usually converted to currency - as in, how much money would you trade something away for? Utility is also different from person to person. Let's return to the carnival game example, where the player had sunk $8 into the game, victory was sure at $15, and the prize was available for purchase for $5 at the store tomorrow. Would the carnival goer rather get the prize now, at the carnival, and give it to their child/sweetheart in this context? Is that difference worth the two additional dollars in cost? Is the warm-and-fuzzy feeling from the charitable donation worth it, especially when the carnival goer has the option to buy the doll for five dollars tomorrow and coldly cut the charity a check for two dollars at the same time? Is your entertainment worth two dollars to you? These questions have subjective answers, but the sunk cost fallacy hinges on their answers. As long as dollars are replaced with utility, the Sunk Cost fallacy is inescapable. Gain/loss of reputation, happiness at acquiring a good in a particular context, secondary effects (such as the charity example), and so on would all roll up into utility.
    • It should be noted that in all these cases the sunk costs are still ignored, the utility of the prize is weighed against the $7 that still must be paid, not against the $8 that has already been sunk, etc.
    • Utility also explains why sometimes our "irrational" choices are not always irrational at all. For example, suppose that in a small American town, a local drug dealer is taken by a customer for about $200 USD, so he goes to the customer's house and assaults him viciously. The possible consequences of the assault could send the dealer to jail for years - so it seems so foolhardy. An armchair economist would look at it and say, "Cut your losses; it's a sunk $200, and risking serious charges over it is stupid." But the drug dealer is part of the black market. He can't go to law enforcement to enforce his property rights over his stuff. A reputation for being easy to con or being unwilling to punish theft could make him easy prey for another criminal. Thus, when weighing the risks against $200 USD and his reputation, and when using the dealer's value system, the choice is no longer irrational. When accounting utility, always look at it from the decision-maker's values and preferences.
  • The relationship between military situations and this fallacy is rarely clear-cut. If you are close to a valuable objective then continued effort may be justified. If this decision was based on one's own sunk costs, it was at best Right for the Wrong Reasons.
    • The fundamental issue is that unlike gambling, in attritional warfare one has to pay the cost regardless of victory or defeat. This then leads to the above problem of committing more troops even when it doesn't help.
    • This was long argued to be the main reason why Russia lost the Russo-Japanese War. By the end of the war Japan was winning militarily, but its economy was stretched to the breaking point, and their mobilization resources were completely depleted, as they had started drafting kids and geezers into the army, with the predictable outcome for troops quality and morale. Some analysts say that had Russia pushed just for a couple of months more, even in the wake of the horrific losses like Tsushima and Mukden, Japan would've sued for peace. On the other hand the Tsar's government had really lousy intelligence (as well as unrest on the home front that needed attention) and didn't know that, so they decided to cut their losses and sued first.
  • When the possible return is so great compared to the possible loss that it is deemed a reasonable risk to take. That's gambling, not fallacious. This works better with non-cumulative risks (like Lotto); otherwise, see pot committed above.
  • Many people often believe that cost overruns on defense projects are an example of this, but usually they are not. For example, while it is true that a new generation of stealth fighters or other extremely expensive stuff can drag on for decades and have cost overruns of 500% or more, it is usually because defense planners want the new system at almost any price. The other reason for cost overruns is what is sometimes called the oversight paradox: If a project costs too much, policy makers may kill the project, meaning that the defense contractor will lose everything they've spent on the project. To protect themselves from this, defense contractors will often rush through the R&D phase as quickly as possible, because a weapons system becomes much harder to cancel once it is in production—too many workers will get laid off, making it politically unpalatable. That means that bugs in the system have to be fixed only after the system is already in production, which is much more expensive, leading to cost overruns. When a weapons system is secret, the lack of oversight often means that the designers can take their time, avoiding problems once the system goes into production; the U-2 spy plane, to give a famous example, came in on time and under budget, thanks in no small part to the lack of Congressional oversight.
    • In addition, even a new project would have similar cost overruns and problems that would arise. It's better the devil you know than the one you don't, and this is another reason why sometimes what seems like throwing good money after bad is really the best option.
    • The Hubble Space Telescope is a good example of why additional costs after the fact aren't necessarily this trope. When the telescope was first deployed, there was a major flaw in the primary optical mirror; it has been ground to the wrong curve, making the images it sent back very blurry, and rendering the primary purpose of the telescope (clearer images outside Earth's atmosphere) mostly pointless. Some people quickly wrote off Hubble as a failure at that point, so when a mission to send up a new module to correct for the bad optics was announced, these people slammed NASA for throwing more money into a telescope they thought would never work properly. The fact was, NASA knew exactly what was needed to correct for the issue, thanks to documentation of the process used to grind the mirror, which was ground very precisely to the wrong shape. Long story short, the corrective module was sent up to (and installed in) the telescope, and afterwards, new images from Hubble were incredibly clear, as had been envisioned at the start of the project. The money spent to correct the optical issues (a relatively small amount compared to the overall cost of the telescope) proved to be money well-spent...and more to the point, the repair cost a fraction of the money that building a completely new telescope would have, for most of the same benefit (one of the telescope's original modules had to be removed for the corrective module to be installed). Hubble is still in service—and providing great imagery—as of 2016.
  • Being statistically "pot-committed" in poker is not this fallacy: sometimes, the pot is so big relative to the cost of calling that the strategically correct choice is to call even when your odds are slim (but non-zero). note  This is more likely to happen in limit games; in non-limit games, players that already have a strong hand are likely to size their bets such that it no longer makes mathematical sense for players on long-shot draws to call. It is part of the strategy of tournament poker to recognize situations where one's stack is low enough with respect to the size the pot is likely to reach that it is easy to become pot-committed. At that point, one is advised to either give up before becoming pot-committed, or shove all-in early so as to increase one's odds of winning by making it more likely that the other player(s) will fold. It is a further part of the strategy to recognize when other players are in that situation and not enter pots with them if unprepared to play for all their chips.
  • An "ideal" option that is not economically viable in your particular situation is just that: not viable, and, therefore, should not factor into the equation. Going back to the car example, stringing by on ad-hoc repairs may be the only viable alternative if the owner just doesn't have the money to buy a new car or the credit to get a loan that would cost less than the repair costs for the period of that loan (or if other means of getting around, e.g. public transportation, biking, or walking, are also not on the table). Previous repair costs are sunk, but it's also no use in bemoaning what you could have afforded if you hadn't paid for those; all that matters is what you can afford now.
  • Repairing something when buying new would cost less can be a conscious decision to save resources and avoid waste.