Truth Bounties

The Truth Market is an interesting combination of prediction markets, bounty hunting and crowd funding that aims to separate the wheat from the chaff of truth claims. Here is how it works:

You hear a statement that you think is bogus (or you hear the denial of something that you think is true). You open a challenge in which you offer to pay a truth bounty of $x if someone can prove that the bogus statement is true (or prove false the statement that you think is true). Other people can join your challenge, adding to the bounty. If the total bounty exceed a significant threshold the challenge goes live.

Once a challenge is live, anyone can earn the bounty if their evidence for or against the claim meets the standards of a neutral, professional, scientifically trained group of adjudicators (provided by TruthMarket). If within a given time-frame no one wins the bounty the bounties are returned to the contributors minus 20% which goes to the initial sponsor of the challenge. The initial sponsor can now also trumpet that despite significant cash no one was able to prove the bogus claim (or refute the true claim).

Thus, there are incentives to offer challenges, incentives to answer challenges, and incentives to pay attention to the results. TruthMarket has some serious people on its management team and advisory board. There are already challenges about global warming, cell phones, defensive gun use and other issues.

Will the Truth Market work? In order to work, TruthMarket will need a track record of significant money bounties and adjudicated claims. As of yet, I don’t see many (any? the site is unclear although this is the most important part of the process).

Most important, people have to regard winning a bounty and the failure to win a bounty as informative. My experience, however, is that the people who regard betting as informative are already rational and well-informed about other issues so the bounty isn’t necessary to prosecute the truth claim.

The Amazing Randi’s one million dollar prize for “evidence of any paranormal, supernatural, or occult power or event,” was first offered in 1964 but has never been claimed. In theory, that tells us a lot. In practice, the failure of the prize to be won does not seem to have changed many people’s beliefs.


Is hand-picked pool of experts model scale-able for purposes of picking "truth" ? Isn't it similar to the patent office model?

On the other hand, I see this as an entirely scale-able model for TruthMarket. There is possibility of a self selection of people who actually pay the bounty, in that they might be more trusting of expert judgments, even when it is unclear whether the expert's field of expertise is relevant, or who the expert panel was (for independent verification of expert bias)

Meta-chalenge for the TruthMarket: verify the efficient market hypothesis.

The proper meta-challenge is more like --

An expert panel is a superior method of determining the truth of contested claims, compared to a panel of random people (e.g., a jury), a statistical sampling of a broad number of people (e.g., a poll), or a vote by self-selected people (e.g., an election).


As I recall, the Amazing Randi has a long and acrimonious history of fighting and arguing with people who have attempted to claim the prize, then argued that the conditions of the test were unfair, the judges biased, the results inconclusive...

For Randi, that's not a problem, because he was planning to argue with those people anyway, and the prize gave them an incentive to submit to his testing conditions. But, the supply of fakers being much larger than the supply of the real thing, if you have any kind of money at stake, you should plan on endless petty bickering with people who aren't happy about your judging standards.

There are numerous studies on the defensive use of guns ranging from 100k to 2Million. I wonder why $4k would produce better truth. Maybe it is just a free-lance talking points market.

The 100k to 2Million fund wasn't for the researcher to take home; either way the conclusion.

From my very brief perusal of the site, it is useless. Most challenges are phased in an unfalsifiable manner such as "Cell phones are safe". I Agee with Andrews assessment.

Once a challenge is live, anyone can earn the bounty if their evidence for or against the claim meets the standards of a neutral, professional, scientifically trained group of adjudicators (provided by TruthMarket).

A new target group for campaign cash.

Are people going to accept the authority of the "neutral, professional, scientifically trained group of adjudicators"?

The key problem is "experts". There are true claims like "There is no God" that experts might be very unwilling to accept publicly, and experts might have ton of other biases.

With normal prediction markets you have some verifiability - and even then there are many dubious cases - intrade market for "Gaddafi no longer leader of Libya by " traded at ~65% weeks after fall of Tripoli when he was on the run since people estimated ~35% chance that intrade will misjudge the market unless he's captured.

Intrade judged this one correctly, but even Intrade traders don't trust them much. What hope is there is "experts" without objective verification?

'There are true claims like “There is no God”'

Got some evidence to back up your true claim?


"Bogus Statement: an allegedly neutral, professional, scientifically trained group of adjudicators selected by a website nobody had heard of last year is an accurate and reliable means of determining the truth of a statement".

Would $5K be enough, do you think?

I say the Truth Market is rigged.

Great concept, but the TM website is a cluttered mess

Sorry, but the Lie Market has been operational and successful for thousands of years. There is far more money to be gained from lies than truth, especially relative to the earning capacity of the liars when equipped only with truth. Also, a pithy lie on a bumper sticker might take volumes of books to demonstrate its falsehood much less alter people's beliefs about it.

The Millennium Prizes are proof that even things strongly believed to be true cannot be proven true despite substantial promises of fame and reward.

How well did the $25 million reward work in finding bin Laden? We aren't even sure the person who gave us information is enjoying his reward. We're not even certain of his death.

There are Holocaust deniers, doubters about Armenian genocide, and conspiracies about who killed JFK.

The best selling books and movies are fiction.

People cannot be convinced of the truth when their livelihoods depend upon them not believing it. Humans love lies.

Destruction has always been easier than creation and always will be. Reproduction can't outrun a meteor, a supernova, or a magnetar pulse, at least not yet.

George Orwell was a prophet.

Markets for truth are incomplete.

I can see the incentive for the initial sponsor (20%).

But what's in it for the rest of the contributors? Why would they join at all?

Yep, it's retarded, just like futarchy.

As soon as human experts are required, truth becomes at least partly opinion, as strict neutrality is impossible especially when required for any complex question with an overwhelming amount of information.

I think there will be a much better model one day based on search engines (no doubt called "Google Truth" :-). The idea is that information sources can be rated by quality (for example, how often they've been "right" in the past, how many other information sources link to them, etc.), each information source may have an "opinion" on various subjects that can be inferred by an intelligent web robot (including links to supporting information which also has quality metrics), and that the quality of such opinions can be aggregated to form a statistical truth (e.g. "Cell phone are safe? True: 82.67%).

I think Randi's bet tell us more about him than it does about objective truth.

Note: I'm certainly not endorsing any sort of magical whatsits.

This TruthMarket thing is retarded for approximately the same reasons that futarchy is retarded (google it). It would insult the idea of truth, by prostituting it to the highest bidder, were not the addition of an insignificant (browse the gratifyingly short list of their claims) NPR-supported PR project to the insults already heaped on the idea of truth altogether insignificant before the combined efforts of the last century.

Are you referring to Mencius Moldbug. I remember him debating Robin Hanson about futarchy, and David Friedman winning.

> In practice, the failure of the prize to be won does not seem to have changed many people’s beliefs.
This tells us a lot too. I even believe this lesson is more valuable than the fact that the prize was never claimed.

I like the way they abbreviate a true statement to 'TS', and a bogus statement to... you guessed it, 'BS'.

Also, I think there are few websites that encapsulate what I would call the 'American Business Publishing Design Aesthetic' as much as this one. It's all about the CAPITAL LETTERS(TM).

Jamie, I am totally agree with your thoughts.

Comments for this post are closed