Augur: a Decentralized Oracle and Prediction Market Platform

The five-minute read

One action, walked through

1. A market is created with a question, an end time, and a designated reporter.
2. Traders buy and sell shares in outcomes. The price is the market's probability only if the resolution process is trusted.
3. After the end time, the reporter submits an outcome with a bond.
4. During the dispute window, anyone can challenge with a higher bond. Further rounds raise the amount.
5. If the dispute reaches a fork, token holders migrate to the universe they believe, and the market resolves in the universe that retains participation.

The argument, unpacked

What has to be true

• Some set of holders will dispute a false report before the window ends.
• The question is specified tightly enough that honest reporters agree. Ambiguity forks honest people.
• The cost of a fork is real to the attacker and survivable for the system.
• Traders understand that a share pays out only after this process, not when the newspaper headline appears.

What happened after the paper

Augur launched on Ethereum, saw thin liquidity, and later revised the token and the market design. The dispute ladder and the fork remain the idea worth teaching. A live prediction market that uses a committee or a single resolver is a different product and should not borrow Augur's security paragraph.

What to check before you use the idea

• Who makes the first report, and how large is their bond?
• What is the dispute window relative to how long a lie would be profitable?
• Is the question unambiguous on the day it resolves, not only on the day it was written?
• Does this market actually fork a valuable token, or does it stop at an admin?

Terms

Designated reporter

The party expected to submit the first bonded outcome after the market ends.

Dispute round

A challenge that raises the bond and reopens the outcome.

Fork

A split of the token into universes, one per contested outcome, used when the ladder reaches the top.

Universe

The branch of the system that holders choose to keep using after a fork. Resolution follows that choice.

The problem the paper names

A prediction market is only as honest as the oracle that says what happened. If a single reporter can settle the market, the market prices the reporter. Augur's paper spreads reporting across holders and makes a false report contestable by people who post a bond.

What the design proposes

• Markets define an outcome space and a resolution time.
• Reporters stake on an outcome. Disputers can escalate by staking more, up a ladder the paper specifies.
• Forking the reporting universe is the final backstop the paper is willing to name.

How the mechanism is specified

• The economic claim is that a liar must outbid everyone who prefers a true resolution, at each rung.
• Market prices are then a public estimate conditional on that oracle actually working.
• A forked universe is a social and technical event. The paper treats it as part of security, not as an embarrassment.

What this page does not treat as proven

• A prediction market can be illegal to offer in a given jurisdiction. The paper is not legal advice and this page is not an invitation to list markets.
• Thin dispute markets fail closed or fail captured. The ladder only helps if someone climbs it.
• Later Augur versions changed mechanics. Cite the version.

Why a venture studio still reads it

The pattern we care about is escalatable truth: a reported fact that a named bond can challenge, with a last resort that is explicit. That pattern shows up in oracles, in governance, and in agent mandates. Augur writes it down.