Augur is live

Augur is finally live.

The decentralized platform for betting on real-world predictions was one of the first applications built on top of the ethereum blockchain, and its creators  sold “reputation” (REP) tokens for over $5 million in 2015 – a time when few were talking about “ICOs” or “utility coins.” A public beta version of the platform came out the following year, and its team published a revised version of its white paper in January.

Now, the Forecast Foundation, the not-for-profit behind Augur’s development, has announced the launch of the long-awaited platform, which was accompanied by the release of the final version of the Augur application as open-source software.


Augur allows participants to bet on anything.

As long as the outcome can be verified in the real world, users can create a prediction market for anything from ether’s price, an election in Brazil or the outcome of Iceland v. Argentina in the World Cup.

What distinguishes Augur from a traditional betting market is that no single party sits in the middle, meaning that users are likely to pay lower prices.

Removing the centralized intermediary from a betting market presents a problem, however: how to bring dispersed, financially interested parties into agreement about the actual outcome of the predicted event?

In Augur’s system, the creator of a prediction market designates a “reporter” to vet the outcome. This designated entity puts down a deposit of REP tokens, which they lose if they incorrectly report the outcome and other REP holders challenge them. The reporter is compensated through fees.

Day-to-day betting is not done in REP, but in ether, the native token of the ethereum blockchain (though, eventually, the plan is to support other ethereum-based tokens). Users can buy and sell shares in particular predictions, which are priced according to the likelihood the market attaches to each outcome.

Here is the full Coindesk article, here is the white paper, here is their home page.


Is Augur overrated or underrated? Shall we bet on it?


"Day-to-day betting is not done in REP, but in ether, the native token of the ethereum blockchain "

Since your bet has to be paid up front and ethereum isn't a stable currency, your essentially placing a double bet on both the specific topic and on the expected ethereum value fluctuation. Many of the bets will be multi-year and the potential ethereum fluctuation is highly significant.

My reading of the article is that you can get out of the bet early by selling your shares of the bet. As a result, I think we might see implied interest rates. If an event is all but certain, some of the "winners" will be willing to collect less than 100% to get out early.

On the other hand, this would seem a good way to leverage an ethereum bet, i.e. bet long-shot dogs. The guys laying odds on favorites are the ones who stand to get slaughtered.

I bet 500 imaginary internet moneys that I'm the world's biggest cuck.

How about an NGDP prediction market?

As long as you can interest a reliable/trusted enough "reporter" to decide things, sounds like you should be able to run an NGDP prediction market on it, if only as a series of bets based on NGDP on a particular future date.

How about a prediction market in everything? Indeed, many if not most financial assets are purchased today not for the income stream but for the potential capital gains. A prediction market for a particular financial asset (e.g., stock in Facebook) skips the middle man. Of course, the prediction market could predict a falling as well as a rising price (thus capturing the short sale market).

Please provide proof assets are not valued on the basis of expected future cash flows and perceived risk on both an absolute and relative basis.

Stop making stuff up Ray.

I can imagine a lot of conflicts with that reporting system, seeing as people with a financial interest can challenge the official reporter. Has the number or jobs in the US gone up or down since 2016? How about the number of illegal immigrants? Have global temperatures risen or dropped since 1900? How many terrorist-related deaths have there been in France since 2001? The world may not be neat & tidy enough for Augur to work very well.

None of your examples are bets about the future, I think you may be missing a lot of the point. Hard to measure future events will have difficulty finding bettors for both sides.

But they can all be specified to be about the future:
1) The number of illegal immigrants in the US will be X in 2021.
2) The number of terrorism-related deaths in France will be X in 2019.
3) Global temperatures will rise by X degrees by 2030.

These are not hard to measure, but if you consider them hard to measure, you just have to specify a particular source to use when adjudicating the issue.

None of those are bets on future events. And wouldn't fuzzy bets have a hard time finding takers?

This new f%$&*# system...

A tort lawyers' wet dream

'Gentlemen, start your engines!'

Imagine a world where electrons comment on electrons moving electrons around.

That's where I think the idea of fetishishing and celebrating the eating of other living things comes from. A weak, but convincing proof that all this is actually real. If we can consume all the simualtion below us, aren't we the simulator?

After 9-11 there was talk in allowing markets like this in terrorist attacks. The assumption being that someone in the group who knew would be unable to resist making bets and so making the information public.

I think the idea is tasteless but probably worth doing. The problem here is that presumably the betting will be anonymous. No central authority running the system. So what is to prevent bad people using it? Terrorists might perform a public service by leaking details of their attacks. But I can see this being used for corruption, insider trading, and all sort of other bad things. How anonymous and/or secure will be it be?

Yes, there was thought of betting markets post 9-11. Actually, it was Paul Wolfowitz who pushed it, IIRC.

Now, the central idea wasn't that some terrorist would be unable to resist a little profiteering, but rather that "wisdom of crowds with skin in the game" would provide a better risk metric than CIA and NSA assessments which had all sorts of perverse incentives associated with them. Many of the prediction markets weren't to be about terrorism; they had broader economic and political questions.

If you see what Tetlock was to do later, the idea had considerable merit. Unfortunately it got nixed due to bad publicity of "US gov encourages people to make money from terror attacks".

Well, how much do you want to bet that Tetlock's superforecasters flood Augur and make shitload of money out of their superforecasts? In fact, I wonder how much Tetlock would be willing to bet on it?

How does this system differ from The Tote, which was set up in the 1920’s by Winston Churchill? The Tote uses a pool betting system as well. Is it that users can initiate bets?

a) Not traceable (even better than gambling for money laundry) b) lower transaction cost c) you can bet on anything as long as you find a counterparty

It's great for the sake of fun, winning thousand of internet points. However, ether liquidity is near zero. Only a year ago the price went from $300+ to 10 cents in a matter of seconds.

A market flash-crashing on one exchange due to too many stop-loss and margin orders does not equal a liquidity near zero.

That's ridiculous. A complete website and an incoporation in UK or Cyprus could have been done by a single person in less than a week.

Yes, but a website would not have a market cap of 300M USD :

A website would be vulnerable to regulation. This is more or less impossible to shut down.

Also to censorship by the owners and embezzlement. The latter, you may remember, is what killed Intrade. With Augur, assuming they coded it right (a big assumption), embezzlement by the Augur foundation is impossible.

I'm impressed as many people are trading on Augur this soon after launch as they are, given precedent like the DAO.

Augur has shot to the top of the "other category" in the most popular Dapp list at

It has had a stunning 229 users in the past 24 hours. That's almost as many as CryptoKitties (300 users).

Solve the equilibrium. In the VC financing world, you should get a viable product as soon as possible to attract funds. In the token world, where money is trapped into trusts and tokens, you should get a viable product as delayed as possible to enjoy your salaries and hedge against the likely collapse of the value of the token when the product is out.

Ha - that's what I've always thought about these crypto companies - they have a huge principal-agent problem.

The priors are that almost all of these companies will fail, so the network owners/agents have a huge incentive to sell their holdings slowly to the investors/principals - all while drawing a nice salary.

Launching quickly to the inevitable lacklustre reception would be downright foolish.

Presumably parties agree in advance on whats acceptible evidence.

The fun starts in the case of vaguely worded outcomes. Betting that Trump gets impeached for example, by the house or by the senate?

Only the House has impeachment power (essentially an indictment). The Senate then tries the accused.

Exactly, but there is certainly a popular informal definition that impeachment = removed from office.

And what if he resigns on the eve of impeachment? Precise wording of bets can become important is all I am saying

I'm going to guess that this will have the same issue that most prediction markets outside of Wall Street have: thinness.

Finally! I look forward to the 'legalized' chess betting.

REP seems problematic. For instance, were Hansen's 1988 predictions accurate? It seems like a really easy question given that in his Senate 1988 testimony (GUradian hosts a PDF) James Hansen graphed some predicted temperature trends based on our emissions path, two of which he called "business as usual" and "draconian emissions cuts". The actual temperatures looked much more like "draconian emissions cuts" over the last 30 years despite the emissions being "business as usual," but his defenders claim his prediction should be judged by re-running his model against the actual concentrations that materialized. SLR is another area where it's not clear how a determination should be made (tide gauges vs satellite, level vs volume). Given what's happened on reddit it would not be surprising if REP allowed all sorts of motivated reasoning.

Pretty sure the emissions substantially exceeded his A (high level) projections. while temperatures were between B and C, closer to C

Great post. Thanks for sharing.

Comments for this post are closed