What if you could bet on whether…
[Balaji Srinivasan] will agree with [“UFOs are aliens”] by [August 1]
[over 50% of people who agreed with x] will also agree with [y] by [date]
That’s what persuasion markets are for.
Persuasion markets are prediction markets, with self-reported opinions as the settlement oracle.
This is important because it could expand the scope of prediction markets to include matters of judgment or interpretation rather than only matters of data.
These self-reported opinions can come from a single person (“Balaji Srinivasan”), a general group (“over 50% of active users”), a filtered group (“users who agreed with [x] before [date]”), or combinations thereof.
Persuasion markets establish several things at once:
A measure of who the public actually trusts to settle questions of public knowledge. The people trusted by the most users to settle their bets, are those who have earned the most trust in the eyes of the user base.
A track record to show how these trusted parties have performed. As oracles settle markets, they establish a track record by which their trustworthiness will be judged in hindsight.
A system for quickly iterating our judgments about who deserves trust. Oracles who showed great judgment in hindsight will be asked to settle more markets, and those who showed poor judgment will be asked to settle fewer.
A lucrative income stream for people who earn public trust. Whoever gets chosen to settle a market, could receive a cut of all the wagers placed. For example, say someone creates this market:
Balaji Srinivasan will agree with [UFOs are aliens] by [August 1, 2023]
If users crowdsource a total of $48,769 in wagers, then Balaji might get 5% ($2,438) just by taking 5 seconds to answer a single YES-NO question.
The most lucrative bet is a public good
The most lucrative bet is to find:
an unpopular conclusion
that has a lot of evidence
and to wager that:
a highly respected person
with a genuinely open mind
will soon believe it.
This is lucrative because 1) most people will bet against you (because it’s an unpopular opinion), yet 2) you have a reasonable chance of changing the oracle’s mind, because the oracle you chose is genuinely open-minded and there’s a lot of evidence. (And it’s a public good because it’s helpful to bring attention and credibility to unpopular truths, and because it’s helpful to identify people who can genuinely accept new evidence.)
The risks are, perhaps the person isn’t as open-minded as you thought, or the evidence isn’t as strong as you thought. Users will need to become good at judging not only evidence, but also intellectual character — who will really change their minds when they see new evidence?
If you can evaluate both 1) unpopular ideas and 2) intellectual character well, you ought to be able to make a lot of money. As a side effect, the conversation may change around that topic when — to everyone’s surprise but yours — the oracle comes out in favor.
Demographics as oracles
After enough opinions have been recorded, it will be easy to filter users based on their beliefs, and on the timing of their beliefs.
Holding certain beliefs at certain times is another way to evaluate intellectual character in order to choose trustworthy oracles.
Consider the third example from the beginning of this article — a bet on whether…
[over 50% of people who agreed with x] will also agree with [y] by [date]
This is an example of using a demographic as an oracle. Instead of resolving based on the judgment of a single person, the “demographic oracle” resolves based on the majority opinion of respondents from a pre-selected group of users, at the closing time.
Aggregating the best evidence
Since the goal is to persuade the oracle, it would be in your best interest to collect the evidence and make it easy to find and understand. This is where the term “persuasion market” finds its most literal expression — it’s literally a market to see whether you can convince someone you trust, of something you think is important.
Therefore, the best evidence for unpopular conclusions may naturally end up in the Comment sections of persuasion markets. These could become a reference point for the majority of online discourse on that topic.
Instant gratification
A simple form of persuasion markets could be implemented pretty much instantly:
Today, someone could make a market on manifold.markets naming a particular person as the oracle. The downside is, there’s no guarantee they’ll follow through and settle the market for you, and it’s only with “play money” so there’s not much of an incentive.
An existing prediction market like polymarket.com could launch persuasion markets as a new kind of market, verifying the oracle’s crypto wallet address to make sure the intended person is the one who settles the market.
But also…
I’m building an app!
I have an idea for an app that I believe could accomplish 80% of what Ideamarket aimed for, with 100x easier UX than Ideamarket ever had. So easy your grandma could use it, and fully web2.
Moreover, it could facilitate persuasion markets extremely well. Can’t promise anything will come to fruition at this point, but we’ll see.
Interested in helping build?
Seeking 2 devs to build the MVP (1 front end, 1 back end)
Donate here to fund the MVP:
ETH & ERC20 — 0x93f9707adb26d98cfc6d73C8840425010AfA968B
We will always support you
Hey Mike, fascinating stuff!
I’m with UMA which is an optimistic oracle. It uses a schellingpoint model with skin in the game to motivate good outcomes.
What’s interesting is that Polymarket went from centralized resolution to decentralized, but that was an upset for some market participants who expected the cultural precedents of prediction markets to continue on, but of course they didn’t—Schellingpoint model has a different jurisprudence.
What you’re suggesting here is a more intentional management of jurisprudence. The times this comes up the most, though, isn’t in cases with disputed evidence so much as ambiguities in the market rules. Things happen in real life that either evade the rules or make one of them really awkward, resulting in a dispute.
You might want to look at some of the historical debates, which you’re right, happen at great length in the discussion format, with UMA. Could be fun to riff on it at some point if you’re keen!