Insights Into Augur — 2018 Year In Review
Features, usage and limitations of Augur’s prediction market protocol
Thank you to the Augur community and the Forecast Foundation for always answering our questions in the community forums.
As the end of the year approaches, some fill their bags in the crypto Christmas sales while others GTFO in classic every man for himself style. While that happens, we have taken the time to analyze Augur’s birth and growth as a decentralized prediction market platform in 2018.
Augur is one of the few Dapps live on the Ethereum main net, where it launched in July of this year. It has been one of the most popular projects in the Ethereum space from the early beginning, being able to maintain most of its momentum since then. Other prediction markets are being built in the ecosystem but none has had the levels of early product usage and community activity that Augur has enjoyed so far.
Almost half a year after its launch, in this post we dive into Augur’s prediction market functionality, provide a number of relevant metrics about its usage throughout these months, and explore its limitations in areas that need improvement.
*Disclaimer: Guesser is building the product that will onboard the non-technical public into Augur markets — launching January 2019.
TL;DR
- Augur is the only decentralized prediction market protocol working on Ethereum main net that lets anyone create and trade event outcomes.
- While participation in some markets is high, with the most popular market gathering ~$3M at stake, a large number of markets aren’t liquid and have barely any activity.
- A community of developers building tools is growing around the protocol, as well as a community of early traders.
- Current user experience is the biggest issue, with non-technical traders facing difficulties to buy and sell outcomes. Curation is needed too.
- Regulatory hurdles, Ethereum’s scalability issues and delays from the Oracle outcome reports are matters to be addressed in the medium to long-term future.
1. Augur’s Prediction Markets
Augur’s markets represent real world events. Markets are binary if only two outcomes are possible, like this market about the price of Ethereum exceeding $200 at the end of 2018 (Yes/No). Others have more than two outcomes, like this one on who will win the Premier League this season (7 options available).
Markets can be created by anyone. To do so, you must provide all the information fields required for a market to be complete and stake an amount of REP for participants in the oracle network to report the outcome correctly once the market ends. Once this process is done, the market is part of the Augur platform and available for traders to invest in its outcomes.
Augur’s smart contracts match opposite orders. The prediction market protocol connects orders from peers that want to participate in a certain event. For you to buy an outcome, there must exist someone else who is willing to invest in the opposite outcome, with the same price per share (which represents the probability of the outcomes) and who is risking an amount of money equal or superior to yours. Example: if you want to buy $30 of a Yes outcome at price 0.5 (therefore doubling your investment if you’re right), there must be someone willing to buy $30 or more of the No outcome at price 0.5. If you wanted to do it at price 0.6, the matching trader would have to be willing to go for a price of 1–0.6=0.4 for the No.
If you buy an outcome that ends up being true, you will profit off the outcome’s final value (worth 1) once the oracle resolves the market. If you had bought an outcome that ended up being false once the event occurs, you will probably lose most of your investment as the oracle will rule it to be worth 0. As all markets are open for trading until the oracle reports, there’s usually active trading before the oracle resolves outcomes to price 1 or 0. If you have a false outcome, you’ll probably be willing to cash out at price ~0.05 instead of 0 so that you lose less money. If you have a true outcome, you might be willing to quickly cash out at ~0.95 instead of waiting for the oracle to provide a resolution. Most markets resolve many of the positions in them before the oracle reports.
Augur’s code is divided into Core, Node and UI. Combined, these produce the Augur App, as explained in its repository:
“Augur App is a lightweight Electron app that bundles the Augur UI and Augur Node together and deploys them locally to your machine. The Augur UI is a reference client used to interact with the Augur protocols core smart contracts on the Ethereum blockchain. Augur Node is a locally-run program that scans the Ethereum blockchain for event logs relevant to Augur, stores them in a database, and serves the respective data to the Augur UI.”
Running the Augur App is the default way to use Augur. Its GitHub repository provides guidance about how to install each release on different operating systems. The App combines the Node, which could be said to act as a registry for markets, orders and transactions, and the UI, which is the interface you interact with when surfing through markets, placing orders, etc.
Augur’s oracle is a set of smart contracts designed to let holders of REP, its native token, agree on the true outcome of a prediction market. Thus, it is a human oracle that relies on a series of dispute rounds in order to achieve consensus trustlessly. Market outcomes are reported by a designated reporter who must hold enough REP tokens to make that first report. If other REP holders disagree with this initial report, a series of dispute rounds can take place if subsequent dispute bonds are filled.
The largest dispute in the oracle system would lead to a fork in the Augur ecosystem. Forks of the REP token based on divergent outcomes create “Universes” for traders to choose in which Universe they will keep interacting with the version of Augur they believe to be correct. Up until now, no dispute has led to a fork, so how this situation would resolve and what issues it would entail are still theoretical premises.
Trading on Augur is mainly done with ETH (you can trade market shares too). This means the platform is subject to the currency’s volatility, which is relevant particularly for long term markets where your ETH might be locked up for months if you don’t want to close your position earlier. DAI will be integrated in Augur v2 (current live version is v1.8.3).
2. Usage of Augur
The following metrics and charts are available at curiousgiraffe.io/augur. Special thanks to Peter for answering our inquires. Data Dec. 11, 2018.
Since the launch of Augur on the Ethereum main net in July 2018:
- 1,635 markets have been created. We can observe an initial excitement around launch date, somewhat stabilizing after that to a range of 1-20 markets created per day.
- 11,825 orders have been created. This metric is, in our opinion, the one that best reflects the users’ will to invest in real world event outcomes. These are orders placed by traders willing to buy shares of a certain outcome.
- 6,331 orders have been filled. This is the number of successful trades that have happened between peers participating in the protocol. The big spike in November relates to the popularity of the US Midterm Elections market.
- 6,119 orders have been cancelled. These are orders that don’t convert into a position for the trader, for causes like no matching opposite order being found at the specified price. They represent 51% of all orders placed.
- 750 markets have finalized, that is, their ending date has passed. Taking a look at the cumulative curve, we can see the rate of markets finalized grows steadily.
- The average duration of finalized markets is 42.5 days, with a minimum duration of 8 days and a maximum duration of 142 days.
- Money at stake in Augur reached $2.96M on November 12. This is money staked in all markets at a certain point in time, its evolution in USD and ETH is shown below. Augur’s term for money at stake is Open Interest.
- Dispute crowdsourcers have been completed 254 times. When deciding on a market’s outcome, reporters can contribute to a dispute crowdsourcer by staking REP on an outcome different than the current tentative outcome for a market. Once these contributions meet the bond required, the tentative winning outcome changes and the next dispute window begins. Thus, tentative outcomes have been successfully disputed 254 times.
- The average market creator fee is 0.0105 ETH, currently ~$0.93. Total fees in Augur markets, called Settlement fees, add the creator fee (set by creator) and the reporting fee (automatically set).
- Usage has been stable at around 30 unique users per day over the last month. The US Midterm Elections market usage rate of late October and early November hasn’t been reached again so far.
2.1. Ecosystem
There exists a growing ecosystem of tools built around Augur, some of which are:
- Guesser, the simplest user interface for Augur prediction markets, letting you bet in the outcome of real world events with 3 clicks.
- Predictions Global, a very useful site that lists Augur markets, lets you sort them in various ways and look at their detailed info.
- Augur Casino is a website that lets you access the Augur App from the browser without having to sync your local node.
- Augur Analytics by Curious Giraffe, a compilation of many relevant live data trends.
- Augur Explorer lists the most recent activity in Augur markets.
- Veil is a financial derivatives trading platform built using Augur and 0x. It has some crypto futures live in their Kovan testnet alpha, and a cool UI.
- Augur Insider is a volatility index for Augur markets.
- Augur Dispute Crowdsourcer lets you check the state of different outcome reporting disputes, organized by rounds.
- Reporters Chat is a forum for conversations around the outcome resolution of markets.
For its part, the Foundation has also been able to get engagement around its content and channels, specially the Discord chat and subreddit. Weekly development update posts are launched, as well as release updates through Twitter threads. The team is usually present at some of the popular Ethereum events and conferences too.
Moreover, some people are building independent automated market makers, and good memes have been popping up lately. Documentation for developers is available at docs.augur.net.
3. Limitations
3.1. UX
Ethereum is still nascent, and its UX very limited. This is a problem all Dapps in the space encounter. Metamask and hardware wallets, for example, still imply quite a tedious process when it comes to spending money. Transactions are slow, and tech like state channels are in experimental stage.
It’s also difficult to provide a good mobile experience. This is essential in other industries, but still very hard to do from the decentralized protocol/app’s perspective. Key pieces of infrastructure are not there yet. Downloading Augur’s desktop client and syncing it for the first time takes around 30 minutes.
Augur’s UI/UX requires technical trading skills. There are two charts provided in each market: a candlestick and a market depth chart. These charts provide professional traders with the information they need to understand a market’s activity, but are not easy to figure out from the perspective of an average user who wants to make money by predicting certain outcomes.
The Buy/Sell experience is confusing. If a market is binary and its outcomes are Yes and No, Augur’s infrastructure is designed to actually have only one outcome (Yes), in which the trader can take a long (Buy) or short (Sell) position.
Selling is specially counter-intuitive. When you sell an outcome, you’re sort of saying “I want to make money if the outcome is false even if I think there’s a x% chance that it will happen”, x being the limit price you provide the order. The cost is then calculated as (1-x)*y, y being the number of shares you’d like to buy, and the UI provides you with a calculated final cost for the order which usually leaves the less skilled traders wondering what amount they should increase or decrease in order to fit the cost to their budget.
Free choice in number of shares and limit price damages the UX. In order to place an order in the previous example market, you can either Buy or Sell the Yes outcome. When you Buy, you must provide:
- Your desired price for the chosen outcome (a 0.5 represents your opinion that the outcome is 50% likely to happen)
- The number of shares you’re willing to buy (1 share = 1 eth)
and make the payment. When you Sell, you fill the same two fields but the calculation explained before happens in order to give you an estimated cost.
As most people don’t buy a round number of shares and the market graphs provided are not so easy to figure out, it’s hard to be precise with your order. This produces wasted time trying to figure out the price and number of shares you want to place to fit your budget, and even once placed, we have seen that 51% of orders are not fulfilled.
As we understand it, this is sort of a Paradox of Choice scenario where many users with less skilled trading profiles would probably be more comfortable actually having less choices to make, less decisions to consider, and a simpler experience where liquidity for orders is easier to achieve.
3.2. Curation
The vast majority of markets in Augur are not liquid. The probability that an order placed on a market gets filled can be somewhat estimated looking at two factors: the spread and the order book depth. Predictions.global, a site that helps organize and sort Augur markets, estimates that less than 100 markets are currently liquid. This is a low percentage of total markets, and we believe it’s an optimistic figure. Thus, placing an order on most of the markets one encounters when searching through the Augur app is useless; a big number of them don’t convert into a position.
Creators can be imprecise when defining a market. If anyone can create markets, anyone can make mistakes (or troll) too. Creator fees can be unfair (e.g. 50%), or market specifications such as ending date or even the wording of the market can be wrong. For example, this market on whether the price of ETH would be over $500 at the end of September had its expiration date on September 4th, way before the end of the month. It gathered thousands of dollars in orders. Even the popular US Midterm Elections market is foreseeing disputes about its correct outcome, as the wording for it was “Which party will control the House after 2018 U.S. Midterm Election?” and the newly elected party doesn’t “control the House” exactly after the election, but weeks later. Since markets can be reported as “Invalid” by the Oracle, this causes some trouble for traders that might have invested in a market, be right and expect profits only to get their initial money back (minus fees) if a market is ruled as invalid.
Market categories are very chaotic. As creators can provide any tag to their markets, there is currently a long list of topics where duplicates are common (Crypto/Cryptocurrency, Sports/Sport), something that makes the sorting of markets in the user-facing client quite a mess.
3.3. Others
Prediction market growth will entail some legal issues. The Forecast Foundation is an open source software development group. It is developing an implementation of a software that lets anyone create and participate in prediction markets in a fully permission-less way. Regulation on crypto and smart contracts is still very young, with bodies such as the SEC or the CFTC just starting to provide guidance in these matters. That said, complying with regulations of one sort or another will probably be a task for the clients building on top of Augur, not for the underlying protocol. We’ll see how all this plays out as the industry matures.
Augur’s underlying protocol, Ethereum, is still experimental technology. Most things in Ethereum are still to be done: the shift towards proof of stake, sharding and scalability issues, state channels… Many things can go wrong and projects built on it will have to deal with upcoming issues. The nature of the smart contract platform also presents some specific risks, like parasitic contracts, which are yet to be solved in Augur. Circle Research has also pointed out to an accumulation of fees due to the cryptocurrency landscape when one wants to try Augur:
“Although the fees market creators and reporters collect are low (1–2%), there are layers of fees users must pay to use the platforms and these fees add up. In Augur, from lowest to highest, these include reporting fees (0.01%), market creator fees (1–2%), Ethereum gas fees (depends on the size of order), and fees for converting fiat to ETH (4% on Coinbase if using debit/1.5% if using ACH). Thus, total fees for trading on Augur range from 3.5% to 9% or more.”
Augur’s oracle works, but competition is heating up. Augur’s oracle is a network of humans reporting on the state of affairs which has been working OK up to date. The first report can happen in less than 3 days, but disputes can go on for up to 60 days. These delays raise the question on whether unmanned data oracles like Witnet might make sense for certain prediction markets that can benefit form instant resolution. Or a general-purpose dispute resolution network like Kleros. Another question that has been debated in the community is whether it makes sense for REP holders to only act as oracles for Augur prediction markets or if it would be positive to open the network to other dispute resolution use cases where token holders could be able to provide value and earn further profits (this would imply relevant upgrades to the code and incentive mechanism).
4. Conclusions
At the moment, there aren’t many relevant prediction market platforms to compare Augur to. There are not even many Dapps with relevant usage on Ethereum. It’s still early days, and this makes it hard to draw relative conclusions about its functionality and usage.
Augur’s basic infrastructure is working, and new releases with upgrades and improvements are being constantly shipped. This is encouraging for developers, who have a protocol they can work with, as well as for early adopters with technical skills that are eager to trade and who have a nascent but functional platform to do so.
Liquidity is a relevant challenge. More orders are currently cancelled than filled, which is disappointing from the user’s perspective, specially when having made several efforts to try a product in the current stage Augur (and the broader Dapp scene) is at. Liquidity shouldn’t be confused with open interest: a market can have good liquidity without a huge amount of money at stake, and vice-versa.
A smart contract protocol must be useful to a very diverse spectrum of people. It is able to do so by providing a set of open source smart contracts that any developer can work with to provide more specific groups of users a highly customized product. Augur Core gives developers the tools to do this.
The Augur UI, present in the Augur App (which is the most popular client for Augur prediction markets) has to be more specific than the protocol by definition. It has to target the best usability for a specific set of users. Although sometimes questioned in the community, we believe the Forecast Foundation’s decision to position their UI towards more skilled traders is actually a smart one. It can serve technical people well for early adoption of the protocol’s prediction markets and although the local running mode is not the most convenient approach, it provides higher censorship resistance.
To achieve the best degree of UX and embrace the average user that is not technically skilled neither a professional trader, curated user interfaces must be built on top of Augur. These probably won’t have the degree of decentralization that the underlying Augur infrastructure has; that’s the natural trade-off for superior usability. If average people are not able to participate in these prediction markets, they will never achieve the relevant levels of liquidity and open interest needed for a technology to go from hundreds to thousands or millions of users.
Our vision at Guesser is a future where anyone can invest in the outcome of any world event, so we are building the product that will open prediction markets to non-technical users. You’ll be able to buy and sell outcomes of Augur events with only 3 clicks. But we’ll dive into that in our next post. In the meantime, you can sign up for early access at guesser.io.
