Winner’s Curse: Negotiation Mistakes to Avoid | Hacker News Reader

Winner’s Curse: Negotiation Mistakes to Avoid

pon.harvard.edu

105 pointsandsoitis66 comments

Doing my best to summarize what I got from this article:

1. In an auction, is possible to over-bid on a thing

2. If you over-bid on a thing and you win, that's not good

3. It's better to not win the auction than it is to win it by bidding higher than you value winning the auction

4. The Big Insight: before you bid, stop to think about whether you're about to bid higher than you value what you're bidding on.

...what am I missing? Why is this interesting? What does this have to do with negotiation? What are the negotiation mistakes to avoid?

You're not missing anything. This article is junk and says nothing that we don't already know about the winner's curse (or you can easily look up anywhere.) It's only a vehicle to sell reports via that popup.

And now you can look it up in this article, written in different words and from the perspective of the individual participating in an auction. That's not junk.

We? I didn’t know this had a name or was formally recognized, and at best I had an intuition for it rather than an explicit understanding of what was going on.

This article solidified the concept for me, and I’m grateful. Why so cynical?

> This article solidified the concept for me, and I’m grateful. Why so cynical?

If it were not posted on HN you wouldn’t have seen it, I’m guessing. As for the article itself, it is a surface-level, paraphrased encyclopedia entry on this observation. It could have provided more value with more than just a jar of coins as an anecdote, perhaps showing how this can apply in face-to-face negotiations.

I think it’s fair to be skeptical of intentions because of low-effort junk out there that exists solely as a lead gen.

OTOH I can now use “Winner’s Curse” as the name of my direct to video screenplay which will appear on Netflix starring that guy from the really good show a couple years back, and maybe with a Bruce Willis cameo

It not only describes the issue it also provides some tools for decision making such as:

a) widening one’s aperture of what is an auction:

Auctions pervade our world, from mergers-and-acquisitions deals and procurement auctions to eBay.

b) ideas to avoid the trap:

- Analyze whether the asset has a common value element. A common value asset, like a jar of coins, has equal value to all bidders. If so, bid with caution.

- Assess your capabilities and compare them with those of other bidders.

- Before placing each bid, pause to consider how you would feel if you won the auction.

Only bid if you have already sold the thing to someone not in the auction. :)

https://en.wikipedia.org/wiki/Winner%27s_curse

If you have alternatives to participating in this auction, perhaps you should abstain. Also related to BATNA: https://en.wikipedia.org/wiki/Best_alternative_to_a_negotiat...

That's only true if items with a defined but unknown at the time of bidding. A piece of art doesn't suffer from the winners curse if you are buying it to enjoy.

Yes it does. There could be another piece of art that you would "enjoy" equally, but pay less for. If no other piece of art will do, then you yourself are equivalent to the pocket buyer.

edit: if you're using "enjoy" to mean "have a private value rather than a shared one," a pocket buyer is a private value. If, instead, your level of enjoyment of the item is around the same level as the rest of the bidders, then there's no reason why the Winner's Curse wouldn't apply.

> If you have alternatives to participating in this auction, perhaps you should abstain.

totally off-topic I know -- it's a great answer to so many dull thought experiments too. Like this one that was going around Twitter recently:

    Everyone responding to this poll chooses between a blue pill or red pill.

      - if > 50% of ppl choose blue pill, everyone lives
      - if not, red pills live and blue pills die

Just walk away! It's so easy!

A core point is the line "Analyze whether the asset has a common value element. A common value asset, like a jar of coins, has equal value to all bidders. If so, bid with caution."

In auctions where the asset does not have equal value to all bidders, the winners curse is less likely to apply. Because if you guess the value of the object higher than all others, that can happen because it is fundamentally worth more to you, rather than because you had the most over-estimation.

Which is why "jar of coins" is a bad example. To a collector, a jar of coins may be worth more than face value, depending on what is in there.

Intuitively, you might think "I'll just bid a few cents less than what the thing is worth to me, then I'll get a fair deal". However:

1. You likely don't know the value exactly, so you might overbid (or underbid)

2. If you overbid, you will probably get the thing and lose money

3. If you underbid, you are less likely to be awarded the thing, so on average the gains (from getting it below value) are lower than the losses (from overpaying).

4. The Big Insight: To still get a good deal on your average auction win, you need to adjust your bid downwards to account for your imperfect knowledge.

Yes. It's not a terribly interesting article but I would describe it from an external perspective as,

"People bid based on a sort of mean value: they don't know how much this thing is worth to them, so they make an uncertain guess, that guess has a mean.

"But, that guess also has a standard deviation. And if you have 100 identical people doing a one-time closed bid on the thing, they all have identical mean valuations but the bid that will win will probably be 2-3 standard deviations over the mean. In these situations, the more wrong you are, the more likely you are to win, and the less you will like what you win. This is the winner's curse.

"You can avoid this by being cognizant of your status as one of the pool of 'identical people', how many people are similar to you, and tempering your bids downward in relation to that. A jar full of money has a similar value to everyone who is in the audience -- this creates the dangerous condition! A unique statement piece from an otherwise unremarkable artist, is going to attract some interested parties and some revulsion: so if you dig it, you are probably in the minority and don't need to worry so much about the winner's curse."

Overbidding has guaranteed downside. Underbidding has non-guaranteed upside. It’s pretty clear that if you’re playing an iterated game then you always underbid. Of course it’s a different story if it’s a unique item that you really want.

Methinks you just restated GP’s points with slightly less clarity

It's just discounting one's bid by the amount of the perceived risk.

For example, we discount the price of a used car by an estimate of how much of its useful life is used up. But that cannot be gauged precisely, hence risk.

We discount the price even further if bought from an individual rather than a used car dealer, because the risk of hidden defects is higher.

Risky investments are priced lower than more secure investments.

Many people say that free markets require perfect information. They don't, not at all. The imperfect information is called "risk" and affects the price accordingly.

Not surprisingly, your explanation is the most coherent of all of them

Another way risk is reduced by paying more is purchasing warranties and insurance.

No, that poster's #3 added the key insight.

Or to put it another way.

Everyone has incomplete information. If you bid more than everyone else it's because one of the following is true.

1. It is more valuable to you than anyone else.

2. Everyone else is mistakenly undervaluing it

3. You are mistakenly overvaluing it

If you don't have inside information, or have a unique reason why it is worth more to you than anyone else. Then 3 is far more likely than 2.

Could you say this is "wisdom of the crowds" working against the bidder?

> 4. The Big Insight: before you bid, stop to think about whether you're about to bid higher than you value what you're bidding on.

How do you define value? How does it consider fungible factors?

----

My wife and I "overbid" on our house. We have absolutely no idea how much we "overpaid". Could be 0%, could be 10%, could be -10%.

If we did it again, we'd take the exact same approach. Winning that bid had little to do with the actual value of that house, other than knowing where to put a value on the bid. It solved a major problem for us - we need to move to a new city where housing in in extremely short supply. Putting that bid in and winning meant saving weeks/months of effort looking, touring, driving out of state, putting bids in, holding our breath, etc, etc.

In an auction, the winner will always give an unnecessary concession (whether that's excess cash or some other contractual term). That's simply the nature of competition; especially with asymmetrical knowledge.

It seems like you're using a different definition of overbid here. You didn't bid higher than what you value the house at, presumably, otherwise you wouldn't do it again.

The article clearly says that the phenomenon applies only to auctions where the final value is approximately the same to everyone.

You describe a situation where a quick purchase is very valuable to you, so getting the sale is worth a higher bid to you than to someone who could take their time getting a good price, such as an investment buyer.

So the phenomenon is much weaker or irrelevant in your case because all buyers are not estimating the same absolute value to them.

> How do you define value?

The buyer defines it as he sees fit, so does the seller. No sale happens unless they both agree on the value.

> What does this have to do with negotiation?

You can negotiate for a contract job such that it costs you more to do the job than your negotiated price, or otherwise isn't worth it.

I agree its not the most insightful article

I haven't read the article, but "The Winner's Curse" is an economics concept I'm familiar with.

Basically, The Big Insight is that every bidder has an estimate of the value of something, and the true value might differ from these estimates. If you're the winner of a transaction then you have likely overestimated the value. Since you had the highest estimate, you probably overestimated.

But it only arises in certain situations. In practice, it depends on whether all bidders have the same information. If you have more information your estimate might be correct. It also depends on whether the item has different value to different bidders, as it might truly be more valuable to the winner.

https://en.wikipedia.org/wiki/Winner%27s_curse

> Savvy bidders will avoid the winner's curse by bid shading, or placing a bid that is below their ex ante estimation of the value of the item for sale—but equal to their ex post belief about the value of the item, given that they win the auction. The key point is that winning the auction is bad news about the value of the item for the winner. It means that he or she was the most optimistic and, if bidders are correct in their estimations on average, that too much was paid. Therefore savvy bidders revise their ex ante estimations downwards to take account of this effect.

It sounds like they don't win anything either. Ultimately the expected value of the bid devolves into a binary outcome: win or loss. This strategy seems to guarantee a loss as the optimal bid is never the maximum bid, ergo it will never result in a win. What's the point? Might as well not bid at all to achieve the same outcome.

The article and all the comments I’ve seen really miss the essence of winner’s curse.

Let’s say there’s an auction for a gold bar. Everyone has to use their own scale. Our scales don’t have bias, that’s easy to detect and fix. But, they do have noise. Let’s say +/-1%, all equally likely.

There are 50 people at the auction. The bar is 100g. The two highest readings are going to probably be near 101g, and those two are going to bid each other up. The winner will have overpaid. Simulating this, the mean second highest estimate will be 100.96.

The essence of this is that valuations have a randomness to them. It doesn’t have to be a scale. It could be a model or just the ideas of a person. Because of the ways auctions work, the highest valuation gets chosen. So, the noise in your valuations doesn’t cancel out. The winner is cursed by overpaying.

You can correct for this. There are formulas. But there are also things you can use intuitively. This is only for cases where you care about the value of something, like you may resell it. Bidding on a taxi to make it to a wedding doesn’t apply.

Never bid what you think is a fair price, like you might in an open market. You need to be getting discount. The more bidders and the noisier the valuation, the higher the discount. If you’re bidding on something where it could be worth 1/2 or 2x easily and there are 100 others, you need to think you’re getting a screaming deal. Cause the most likely reason you’re the higher bidder is you were unluckily high in your valuation. Again, if you have a reason to bid high (that was my dad’s corvette!) this doesn’t apply.

This is, unfortunately, more frequent problem than many people think.

I have seen so many contracts in the corporate world where each side wants to get as much as they can for themselves. What they should be doing instead is figuring out how to make both sides of the contract happy. They should be thinking and answering "If this contract is signed, would both sides be happy and would it result in a lasting and successful relationship?"

The best strategy is to price your services right, make sure the contract is adequately serving both sides and, ideally, that both sides will be getting something from it that will make them happy to continue the relationship.

On the topic of negotiations, I would recommend a book “Never Split the difference” by Chris Voss a former FBI hostage negotiator. He takes you through his own experience of handling various hostage situations, salary negotiations etc.

I enjoyed the interview with him on Lex Fridman’s podcast, too. Some repeat of the book, and some new material.

I recommend the book for the psychological/emotional tips.

Otherwise, it's a pretty poor book:

https://news.ycombinator.com/item?id=27335076

This book has created a whole class of very annoying people who do nothing but repeat "How can you make this deal better for me at no cost to myself" in an endless loop and have no strategy for what to do when the other party isn't willing to negotiate themselves downwards for free.

Both parties can employ that strategy, and there's no sale if they cannot agree. It's self-correcting.

If you find it annoying behavior on the part of the other party, walk away.

"How am I supposed to do that?"

"Well, I've got these hostages here, so I suppose the 'how' is up to you."

People who take hostages aren't very smart, so being poor negotiators is par for the course.

How do I know the hostages are alive?

What else can you do? Err on the side of all the hostages being dead?

That’ll make for some spicy firing + public disgrace + jail time if you’re wrong.

I read all the classic negotiations books and think this one is the only one you need to read. IMO its core idea is NOT "how am I supposed to do that", as other posters say, but the role of empathy and uncovering hidden information in high-stakes (or even regular) negotiations

I was hoping for more. I guess I watched, bid and lost enough eBay auctions to realize that people would bid up the price of a used item well beyond the price of buying it new on any number of retail sites. The thrill of winning was the only explanation I could come up with.

eBay or whatever the "jar of coins" is, I try to figure out what it is worth to me before placing the first bid - and never go beyond that. If I fail to overpay, I'm OK.

Back in the day, there was a certain exciting novelty to eBay, often for people who had little experience with auctions. People could also convince themselves that if everyone else thought something was worth more than they did, maybe they knew something--but most likely they didn't. So, yeah, the thrill of the hunt was certainly part of it.

I suspect some of those are shills.

The problem with eBay auctions is the time limit. It pays to "snipe" them one second before the close. I realized this in the early days of eBay, and within a few months everybody knew it.

eBay could fix this by not closing an auction if there are bids within the last 5 minutes, but they show no inclination to.

I have ZERO online bidding experience but have gone to a few in person (equipment, tools, livestock).

Some advice I received from a general contractor applies to auctions: You never lose money on a contract you don't win.

I've seen FOMO and pissing matches run up bids when simply putting a value on each lot you're after and sticking to it seems best.

Leaving an auction empty handed only feels bad for a few minutes and is easier the next time.

Once upon a time, at least in person, the auction house took there share out of the sale price (the buyer's share) but everyone where I am has charged it as a fee to the buyer for at least 15 years, so consider that when setting your number.

Anecdotes:

I've been in a bidding war, hit my number and stopped, had the exasperated Auctioneer try to goad me, and lost.

I've witnessed unbridled enthusiasm make equipment sell for 80% of the cost to buy new.

Once a bidding frenzy at a fur auction meant the price of squirrel pelts doubled and my colleague made $16K snaring squirrels that year. BTW: At least in the 90's, the People's Liberation Army was the biggest buyer of squirrel fur; I was told they were used to fringe hoods on winter jackets. That may have been as case of get the pelts no matter what.

> At least in the 90's, the People's Liberation Army was the biggest buyer of squirrel fur; I was told they were used to dr hoods on winter jackets. That may have been as case of get the pelts no matter what.

If Mao hadn't gone on his retarded crusade against squirrels, they wouldn't have had to do that lol.

Opening a harvard site, having to close a popup with "No I dont want to win negations..."

Then forcing popups.

Seems like they went all in.

In bond trading a BWIC (bids wanted in competition) or 'bid list' is a list of bonds for sale that a seller would send to a number of banks with a set time for all bids to be in. As a bidder you would be notified of whether or not you won the bonds, and the what the second best price was (the cover). This would allow you to know where you were with respect to the other bidders.

Except back in the old days (pre-crisis) there was no way to keep people honest, and people would routinely lie about the cover and always tell everyone the cover was just a tick below the winning bid, so if someone made an egregiously bad bid, hopefully they'd never realize...

Now all transaction get printed on the TRACE feed, which is probably better for the market.

Reminds me of house buying. Realtors are experts at getting bids and creating a high stakes competition with 0 transparency or accountability. They can easily lie or mislead people (and a lot of them do).

When you place bids or asks in a market where some participants are knowledgable ('informed'), the execution of your transaction indicates a likelihood that your counterparty is one of those informed traders.

This should prompt you to adjust your initial bid or ask price. Knowing this dynamic in advance, you can set your initial bid or ask with the assumption that it will be executed by an informed trader, or assuming an X% chance that the trader is informed. This is part of the reason for the existence of bid/ask spreads.

Winner's curse seems similar. When deciding your bid, you have to decide 'if I will in the auction at this price, will I be happy?'.

This is subtly different from "what's the most I'm willing to pay based on the information I have today?".

I have settled on a strategy of bidding that seems to work when I am actively bidding (this goes only for little stuff not land and such). When I hope the other person bids, then I know I don't want the item at the stated price. I can feel intuitively that I will not want the item if I get it.

Does anyone have a help for the opposite scenario? In this scenario I have not yet bid but the price is quite low and I don't want to bid in case I missed something. The price stays very low and the item is sold. I am left kicking myself for not sticking my neck out there and making a good deal. Maybe I should bid once and see if I like it. The problem is that one bid might get it...

I have extensive experience bidding on eBay auctions, I probably engaged in 10 auctions a month for several years in a row.

The first few times I definitely encountered winners curse. But the real/only solution is having accurate price discovery and having a game plan. I will do research to find the current market value and bid in the bottom end of the range. If I miss it, the market is liquid enough such that I can try again and if i hit it, then I am happy because I got it for a low price.

On items with lower liquidity and availability I still need to be patient and disciplined with my bidding. This is the only way to do online auction bidding.

Oh and also sniping in the last 30 seconds.

> Oh and also sniping in the last 30 seconds.

I really don't know why eBay hasn't "solved" this.

In real world auctions, bidding continues until no one bids. This eliminates sniping as a tactic.

My favorite auction site was ubid.com, which had this policy. There always needed to be a 5 minute period of no bids for the auction to end. This way, you can comfortably tell the automated system your max bid and not worry about snipers.

This is basically the adverse selection concept which is explained incredibly well in Agustin Lebron's book (ACX review - https://astralcodexten.substack.com/p/your-book-review-the-l...) but without as much depth or nuance.

Price discovery in an auction is great for the market, not necessarily for the bidder with incomplete info. This article feels like a long intro to an article that never actually starts.

NB: No offence to the author if you read this! Criticism is probably just because it was info I already knew.

I think that the point that should have made the article but did not:

- in the context of negotiations, don't forget that sometimes it is better to withdraw than trying to win at any costs just to win.

- and you should remember that you don't negotiate just "to win" but because you wanted a positive income.

I’m currently building a set of tools to teach negotiation skills through interactive scenarios, with a hope to focus on enterprise sales, and then eventually using machine learning to analyse actual negotiations. If that’s something you have a need for right now (esp if you work in enterprise sales), do feel free to drop me an email at pete@negoat.io

Like the infamous UK 3G radio spectrum auction that both raised huge amounts of money for the Treasury and led to the winners suffering massive loses.

http://www.gsmhistory.com/3g_auction/#3g%20auction

I think what this article misses is that 'value' is a personal/subjective thing. Value is not the same as price. To me, the value of a house might be higher than to you (say because I am more desperate to move out of a tiny apartment). This does not hold for the jar of coins example in the article of course.

Value is the price you would buy/sell it for. There is no other useful definition of value.

I think this applies to more than just auctions. Bidding to take on a contract or a responsibility, that turns out to be more than what you expected. Or overselling yourself and going through great pains to reach a position with very few worthwhile benefits.

Thinking of the scene in Succession now 'Congratulations on saying the biggest number!'

https://www.youtube.com/watch?v=1E_L87IJnw8

Auction Theory [0] might be of interest.

[0] https://en.wikipedia.org/wiki/Auction_theory

This is just "Buyer's Remorse"

Value and price are wonderful, but something is worth what people are willing to pay for it.

The negotiation is won before the actual bid by having a better cost / payoff structure going in.

If all bidding processes followed rational rules.

Reply on news.ycombinator.com