The sunk cost fallacy | Hacker News Reader

The sunk cost fallacy

bbc.com

An interesting twist: sometimes it may be worth finishing purely for the psychological boost of confidence, or other similar meta, 2nd order effects.

I don't have a good algorithm for knowing when this is the case, but especially for smaller individual-level projects, I think it is probably not that uncommon. It also seems highly context dependent -- e.g. if one hasn't had a win in a while, it should be a larger consideration.

Framed another way: first learn how to finish projects, then learn how to finish the right project.

And of course, there is the problem that all outcomes and evaluations of worthiness have at least some uncertainty, if not significant amounts.

The sunk-cost fallacy is a useful tool, but it isn't always easy to apply in practice.

Those effects are compatible with a proper understanding of the sunk cost fallacy. The 2nd order effects are future benefits that can make the project worth completing. The point is that it's the magnitude of the benefits you expect to get that matters and not the amount of resources you've already spent.

Totally. But I think 2nd order effects are often underappreciated, and the difficulty of evaluating them does take some of the "oomph" out of the sunk cost fallacy.

> Ever gone to the cinema and stayed to the end of a film you actually loathed 10 minutes in?

I think the author is assuming that what preceded will therefore follow. Hate the first 10min? The rest of the movie will be bad. Lose $100 gambling? You'll keep losing.

I think people evaluate more based on percentages and alternatives. What's else an I going to do if I leave the movie early? What are the odds of me enjoying that activity versus the chance the movie night turn around or the value I'll get from being able to knowledgeably talk about the film with friends.

Or, did I lose the money gambling from making bad bets or is it just the natural up and down? For example, if you sell a stock every time it goes down in value, you'll be pretty broke.

I think people choose to carry on with diminishing or unlikely returns if the likelihood and magnitude of the turnaround is sufficiently more attractive than stopping. That reasoning makes perfect sense as long as you're evaluating your options accurately.

I was always taught not to consider sunk costs in your decisions, just know where you stand in the present and think about what you want in the future. What happened in the past can't be undone.

The author understands the fallacy just fine. As do you. The flaw is your assumption in paragraph 5. People give the sunk costs greater weight than you, and so you seem to be assuming this fallacy doesn’t really exist.

They stick with a bad movie because they spent the ticket money. Not because they expect the actors to have received lessons and coaching that makes act 3 an Oscar contender. Dinner with their friends would be a better return (for the evening) but they get caught up in that $10 they spent and the “loss” of that if they walk out.

>Not because they expect the actors to have received lessons and coaching that makes act 3 an Oscar contender.

This is the actor fallacy. Sure, there are good and bad actors, but what makes a movie (or play) worthwhile is the script, then the director, then other stuff like cinematography, set decoration, lighting, etc. Actors are actually the most replaceable and least important part of the equation.

Counterintuitive, I know, especially in light of the publicity machines that are at work 24/7. And for goodness' sake, don't mention this to the actors.

So if the plot premise sucks in the first 10 minutes, you might as well bail out then. It ain't gonna get better.

Fair enough. It was one element of the whole thing and I could’ve listed more.

Sorry, I'm a pedant by nature. I was not trying to pick at you. It's something I learned early that most people misunderstand. I'll return to my usual grammar correcting now. :O

I understand the sunk cost fallacy, but I agree that I'm not sure this is the greatest example of it.

Or maybe I relate to films in a completely different way from anyone else.

For example, gaping plot holes or impossible character motivations often turn me off. But I love apparent inconsistencies that are fully explained by the end.

I basically want filmmakers to convince me they put more thought into their film in the months of making it than I could in the hours of watching it.

Or... some of my favorite films are a slow burn that suddenly clicks into an intense ending that surprises you by raising the intensity far above the setup. The Third Man, Hara-kiri, M, Oldboy, The Hidden Blade.

By either of these criteria, it can be incredibly difficult to know whether a movie is good or terrible in the beginning. The extreme best and worst could begin very similarly, but they'll end in completely different places.

A bit like if Croesus and Solon were just film critics.

There's a fine line between persistence and sunken cost fallacy.

Professional Poker players know that when pot odds are 33% for a given hand, they must consistently play that hand at least 3 times for the odds to fall in their favor.

To the Author's point, though, betting on people is less about odds and more about politics... so failing projects are likely to continue failing.

The article never clearly states what they consider the fallacy to be. So it's maybe not surprising that they seem to actually not understand it. They say:

QUOTE: After you’ve invested £10 million ($13m) in a project, which hasn’t delivered, the case for throwing in a further £5 million is far easier to justify if you only consider returns on £5 million – rather than £15 million. But in reality, of course, you also don’t want to look stupid by abandoning it.

But it is CORRECT to consider only the 5 million additional investment when deciding whether to continue the project. Of course, it's not correct (from the organization's point of view) to consider whether you (the employee) will look stupid if it's abandoned. That would be some sort of fallacy. But what is usually in my experience called the "sunk cost fallacy" is the opposite - insisting that to justify continuing the project, it's necessary for it to pay back the whole 15 million, even though 10 million of that is a sunk cost.

I think they're saying it's correct to only consider the 5 million, then pointed out that you're also biased in that way for other reasons.

"It is never too late to give up."

I know it sounds like a de-motivational poster, but it nicely summarizes this entire concept.

I've tried to explain this to one of my (former) employers, he just couldn't wrap his head around it.

Of course it matters, I paid a fortune for whatever. Yeah, right.

It's really about being here and now, and may be applied to anything in life.

I have to be honest, I think entertainment may be one of those things where to some degree the sunk cost fallacy may not be 100% accurate. I mean yes, I see the reasoning behind this:

> We all do it. Ever gone to the cinema and stayed to the end of a film you actually loathed 10 minutes in – or watched yet another season of what was once your favourite TV show?

But TV shows, films, books, games... they've all got many examples of works which either start out terrible and get much better as they go along or start out fantastic and fall to pieces later down the line. Someone who takes this fallacy seriously all the time wouldn't read/watch Lord of the Rings long enough for the party to leave the shire. They probably wouldn't reach the first dungeon in The Legend of Zelda Twilight Princess (which is sometime after about 2 hours of tutorial content and meandering intros). Hell, TV Tropes has a whole list of works where you have to sit through hours of boring filler before getting to the interesting stuff:

https://tvtropes.org/pmwiki/pmwiki.php/Main/SlowPacedBeginni...

There's also the related idea of a show/work 'growing the beard', where it really picks up later on as the writers/actors/directors/whoever finally get into their stride.

So it's not necessarily a fallacy to sit through a tedious or even bad work under the assumption it'll get better, because in quite a few cases, that's exactly what does happen.

> But TV shows, films, books, games... they've all got many examples of works which either start out terrible and get much better

But for one good example, how many bad examples are there? You are willing to sink your cost by betting on, say, 10% chances that a dud movie is going to turn out to be better.

See also; the sunk-cost-fallacy fallacy.

Naaive accounting based only on financial costs, excluding switching costs, other complementary investments, psychological costs, unbooked benefits, etc., may justify at least dome behaviours typically branded as SCF.

https://xebia.com/blog/sunk-cost-fallacy-fallacy/

https://modleft.blogspot.com/2005/09/sunk-cost-fallacy-falla...

https://econundead.com/2015/09/20/the-sunk-cost-fallacy-fall...

http://www.gwern.net/Sunk-cost

I looked at the first link, and it has little to do with the sunk cost fallacy. If you're not factoring in the cost to switch, you are simply making poor estimates for your costs.

I have a business now that I've sunk $100K into. I have an option to abandon it and move across the country for a salaried job. I need to estimate whether I'll have more money with the job vs my business 5 years from now. If I estimate it wrong by not taking into account the cost to move, or the COL of the new city, it has nothing to do with the sunk cost fallacy.

For the 2nd link, I've never heard an argument using sunk cost that would say one should buy the jumper. This seems like a strawman.

I'll admit that this was a quick grab at some references. I've seen several better, though Gwern's essay is pretty good. I don't have the others readily at hand.

The failure to account for switching or information costs is fairly common among misuses of SCF, and looking as to why that is may be fruitful.

SCF is not completely wrong, but analyses claiming it are very frequently incomplete.

I first read about this when I was a teenager, and frequently think about it when making decisions, but have never once found that being aware of it has been helpful. In the face of uncertainty, you never know what is going to happen. I have never experienced a situation where I have poured money into something and it became "obvious" (for a rational agent) that i could do just as well by abandoning that previous effort.

I would be very interested in hearing other's experiences.

Very useful in business. It's tempting to agree with, "Hey, that $60M got us almost all the way there; for only 6M more we can finish!" But the real question is "Shall we spend $6M for a product that we think will generate $3M/year in years 3-5 after launch?"

Or if you prefer, "We've spent two years building the test infrastructure for our legacy product, and are within a quarter of finishing. Looking forward, will we generate enough support revenue to justify investing 10 people for a quarter (plus documentation)? Or have most of our customers migrated by now?"

> Very useful in business. It's tempting to agree with, "Hey, that $60M got us almost all the way there; for only 6M more we can finish!" But the real question is "Shall we spend $6M for a product that we think will generate $3M/year in years 3-5 after launch?"

That's such a lopsided question it's hard to try to connect it to reality, though. If the expected profit is $9M then how did you manage to spend $60M in the first place? Something huge happened that changed the picture, and makes it hard to relate as a simple "sunk cost" example.

That’s precisely the point. You might have launched that project two years ago thinking it wouldn’t take too long and would be a nice little earner for your (then) last generation product. Well instead of six months it took two years and now that legacy product is two generations old and most customers have migrated...but it’s finally almost done and you do still have some customers using it, and gosh, look we’ve already spent $60M; what a shame to throw it all away when for only a little more.... people make decisions this way all the time.

These decisions also consume labor. Despite all the work put in, is it really worth finishing that refactoring or should you just abandon it all?

> That’s precisely the point.

No, I don't think you're getting my point.

You made it such a huge gap, and not explained it, that it made me unable to properly comprehend the scenario. The psychology of sunk cost is influenced by how we got here, and you threw out such an extreme example that it overwhelmed my attempts to get into that somewhat-subtle psychology.

So if you're saying that the product is already obsolete, and has lost more than 90 percent of possible revenue... that's not really a typical scenario, where you still expect to make a pile of money like in the original plan, you just need to get that last.. bit.. done. And in that scenario the sunk cost fallacy is probably not in your top ten problems either, even if you do listen to it.

>I have never experienced a situation where I have poured money into something and it became "obvious" (for a rational agent) that i could do just as well by abandoning that previous effort.

I think you're not understanding the fallacy. The whole point is that the amount already invested is irrelevant. It's not about whether the choice is easy or the outcomes clear. You've put in a lot of money and 3 years into a project. You're wondering "Should I quit and pursue X instead, or stick to it?"

The fallacy won't tell you what to do. It is telling you that whatever you decide, the fact that you put in a lot of money and 3 years should not be a factor in deciding.

There is a (long) argument here (http://www.gwern.net/Sunk-cost) that people tend to over-use the concept, since it's probably the most know logical fallacy.

Sometimes it can be hard to tell it is a sunk cost that’s the problem

Sunk cost refers to whatever resources have already been spent. Where it becomes a fallacy is thinking along the lines of But I’ve already invested this much money or You know, I’ve been working on it for all these years, so I CAN’T quit now.

What we’ve paid into it is gone — a sunk cost — and what matters is how much additional investment from where we stand now will be required to complete the goal and whether that cost will be worth it when considered against other opportunities for the same time, money, and attention.

In some sense, we start at even every day.

The cost has always sunk. The fallacy is assigning it value.

In the broadest sense the fallacy is about not wanting to waste past effort. Whereas in reality 'wasting' past effort doesn't cost anything but wasting future effort does.

Now it can be hard to tell to what extent your past effort actually brought your goal closer, so it can be hard to judge if continuing effort may or may not be worth it, but if you're arguing it is worth continuing because you don't want to waste past effort, and not because the goal is worth the remaining effort required, then the reasoning is fallacious.

I'm always wondering if this is actually a fallacy.

You buy theater tickets. I say, "give them to me." If you don't, is it because you want to see the play, or the sunk cost fallacy? Does the fact that it is icy out and dangerous make a difference?

The key to the theater tickets example is that you should treat tickets you've bought the same as if you got them for free. Replay the scenario but instead of buying the tickets everyone at my office got free movie vouchers. If I'd give you the free voucher but not the ticket I bought, that's indicative of the sunk cost fallacy.

Better examples usually make the alternative better than the thing you already paid for. Giving you my tickets isn't much of a benefit to me, but icy roads can make a difference. If conditions are too dangerous to drive to a free movie showing, I shouldn't drive to use the tickets I already paid for either.

It isn't the same to keep walking to the shop that you know is closed and will got no benefit out of it, than to keep spending on high speed rail or an IT project where you will end up with a benefit.

My personal experience is that you can only determine whether or not project should be continually pursued is in hindsight. Knowing ahead of time, this decision is mostly based on dogged persistence and sometimes a little insanity.

The F35 jet plane. Enough said.

It's not a failure if it's purpose is to cost a lot of money.

Not a lot,trillions!! The cost is comparable(if not exceeding) the interstate highway project.

Only if you miss what the purpose of the F35 is. If you are silly enough to think that it is about creating an aircraft then it looks monumentally stupid. If you see it for what it really is (a massive heist of the tax payers) then it is a work of genius.

Even then,it's not like it's some massive conspiracy. There are lawmakers and pentagon officials that would habe stopped it were it not for sunken cost fallacy.

Unless of course you actually believe every single person in washington has sold out his country for money.

A lot of people have tried to stop the F35 project so far without any success. Just because a program is continuing does not mean it has not faced opposition.

there is a pretty fine line between persistence and falling into this trap. In hindsight it's clear but while you are in the situation it's hard to tell if you should call it quits or not.

Does google have the opposite of a sunk cost fallacy?

Where was the trick?

At the end of the article

Thanks, Beeb, now I'm depressed -- turns out my whole reason for existence is based on a fallacy...

But it wasn't _your_ fallacy.

Not sure that there’s such thing. I mean if you’re not persistent enough you won’t know if there’s a value to it. And if you are persistent enough you won’t know what the alternative would have been and if that would have been better than you are right now.

Reply on news.ycombinator.com