Four Prediction Market Platforms Compared
A review of Augur, Gnosis, Stox and Hivemind
Prediction markets allow participants to speculate on the outcomes of future events, whether political elections, sports matches or quarterly business reports. Just like traditional gambling in sports or race horses, those who correctly forecast the outcome correctly win money. There is currently a new generation of prediction market platforms being developed which use blockchain technology as an integral part of the design.
According to their creators, these new platforms will be more secure, faster, more transparent, cheaper and more accessible than previous ones. Additionally, some of them claim that they won’t be subject to laws pertaining to online gambling because they will be running on a decentralized network, like Ethereum or Bitcoin. However, these claims need to be viewed critically. Running an online gambling site on a public blockchain and calling it a “dApp” doesn’t mean it will get to skirt all regulation. That said, this article will look at four projects under development: Augur, Gnosis, Stox and Hivemind.
Introduction to Prediction Markets
Though the technology of these platforms is new, prediction markets have been used for centuries in various forms. However, adoption has been limited by technical challenges, geographic constraints and governmental regulation. The emergence of the internet saw the creation of the first truly global prediction markets, with the first online betting exchanges and political betting markets coming into existence during the 1980’s and 90s. Though some created peer-to-peer systems which allowed participants to bet against each other, they were generally run by a centralized organization or company who chose what events could be traded, calculated odds, made judgements on outcomes and handled payment.
Most governments see online prediction markets as either gambling or options trading and heavily regulate them. As a result, today they are only available in certain jurisdictions with limited types of events. Some of the largest commercial ones are Betfair, BETDAQ, and Smarkets, which are primarily used for sports betting, horse racing and other forms of gambling. Prediction markets affiliated with universities, such as the Iowa Electronic Markets or PredictIt, generally focus on political events. Most exchanges earn revenue by charging fees that are calculated as a percentage of net winnings for each customer on each event.
The story of the Intrade is illustrative of what happens when an online betting exchange runs afoul of government regulation. Founded in 1999, Intrade gained widespread media attention for the accuracy of the political forecasts on its site during the 2008 and 2012 U.S. presidential elections. But in 2012, Intrade was sued by the U.S. federal government for unregulated trading in commodities, and subsequently excluded all U.S. residents from accessing the site. Following this, Intrade saw a dramatic decrease in trading volumes, and closed its doors in 2013.
The creation of Bitcoin saw the rise of new betting exchanges that accept cryptocurrencies as payment, such as Fairlay. Decentralized payment systems allow these operators to get around the banking and other regulatory hurdles. Though they allow anyone in the world to participate in prediction markets, crypto betting exchanges are generally unregulated, and participants must trust that the operator is not defrauding them.
How Prediction Market Platforms Work
The new generation of prediction market platforms are using trustless, public blockchains, such as Ethereum and Bitcoin, to allow for, in theory, greater transparency and remove the need for a central operator. These public networks also could allow the operators to avoid governmental regulations. Though each of the new projects is different, there are some similarities in how they operate.
Anyone on a decentralized platform can create a prediction event for a fee. This market creator will subsequently earn fees from participants who trade in the event. Each time an event is created, two or more “outcome tokens” are created. If, for example, the event is a simple binary question, such as “Political Party X will come in first place during the election in May,” there will be two outcome tokens created, one for a positive outcome (Political Party X comes in first place) and the other negative (Political Party X doesn’t come in first place).
These outcome tokens, which are like financial securities, can be freely traded on the platform’s marketplace at any time after market creation. The price of the share reflects market sentiment about likelihood of an event occurring. For example, if a positive outcome token sells for .55 collateral tokens, this implies that there is a 55% chance that the outcome will occur. The share’s value will rise and fall depending on new information and trading.
If, at the end of the event, the share holds the correct prediction, then the buyer can redeem it for 1 collateral token, such as Ether, Bitcoin, or another ERC20 token. If the share predicts the wrong outcome, it will be worth nothing. Generally, the platforms support three types of prediction events: binary, where the outcomes are yes or no; categorical, where there are multiple discrete outcomes; and scalar, where there is a lower and upper boundary defined for the event.
To verify the event’s outcome, there are several methods that can be used, with the most common being a reporter or “oracle” designated to confirm the result. Alternatively, the platform could take a survey of other users who didn’t participate in trading the event and ask them to report on event outcomes. This may require some internet research on their part, but there are incentives to increase their participation. If the designated oracle doesn’t report, or there is a dispute about the outcome, then various resolution procedures can be enacted. Each platform sets up their process differently.
Augur, which was founded in 2014 and runs on Ethereum, describes itself as “a trustless, decentralized oracle and prediction market platform.” Augur believes that having liquid, efficient prediction markets would benefit policy makers, the private sector and the public. The company has a very pure vision of their platform and don’t want to control how people use it.
Augur won’t receive any fees or manage any aspect of the trading or operations. The smart contracts that Augur runs on will be totally automated: developers won’t be able to spend the funds held in contracts, control how markets resolve, approve or reject transactions on the network, nor would they be able to modify or cancel existing orders.
The Forecast Foundation, a not-for-profit corporation, is the legal entity behind Augur that is responsible for managing its promotion, development and maintenance.
Augur held its ICO from August to October of 2015. It issued 11 million of its native Reputation (REP) tokens and raised around $5.3 million USD. REP began trading on exchanges for around $1.47 USD and is currently trading at $30 USD. This places it in the top 20 of all altcoins in terms of returns according to TokenData.io. Its REP token currently has a market cap of around $350 million USD.
After the ICO, the tokens were distributed as follows: 8.8 million (80%) of the tokens were sold to the public. The other 2.2 million tokens were distributed between the team/advisors (16%) and to the Forecast Foundation (4%). Since Augur doesn’t take any fees from its users, the founders and team only stand to benefit from Augur if the value of REP tokens increases.
The white paper lays out the operation of Augur. All trading of prediction events is done in Ether. REP must be held only by those who want to create events and or act as reporters/oracles. Users can earn fees from engaging in these activities.
Event creators must post bonds, or stakes, in ETH and REP. These bonds will be returned to the creator if 1) the market resolves properly (meaning the outcome can be determined), and 2) the reporter chosen by the creator reports properly. Market creators earn fees from other users who trade in the event.
Users who want to act as reporters will also have to stake REP. They will be incentivized to report because each staked REP token entitles its holder to an equal portion of Augur’s market fees. Also, if they report a result incorrectly, then they lose their stake. The more REP a reporter owns, and reports correctly with, the more they will earn for their work in keeping the platform secure.
If there is a dispute about the outcome of an event, the platform goes through several rounds of resolution. In extreme cases, the platform will go to forked mode, where trading is halted, and every user can vote to determine on the outcome of the disputed event.
Augur is currently running on the Ethereum test network (Rinkeby) in beta version, which was released in March 2016. The alpha version had previously been released in June 2015.
Augur has been releasing weekly development updates on their blog, with the latest occurring on March 7. In their last update, they announced they just finished the security audits, which were done by an external vendor, Zeppelin. Augur claims all published issues were directly fixed or addressed.
According to the roadmap, completing the security audits were phase one of five. No release date is mentioned, but it seems possible that Augur is aiming to have a final version available in 2018.
Team and Advisors
Augur was founded in 2014 by Jack Peterson and Joey Krug. Peterson received a Ph.D. in Biophysics from the University of California, San Francisco. Krug studied computer science at Pomona College, and is also the co-chief investment officer at Pantera Capital, a crypto investment firm. The total team is around 15 people.
Augur’s advisors include Ethereum founder Vitalik Buterin; Ron Bernstein, the founder of Intrade, one of the first online betting exchanges; Dr. Robin Hanson, an economist from George Mason University; and Elizabeth Stark, co-founder of Lightning Labs.
The markets Augur creates could be viewed as options trading or gambling by various governments, leading them to regulate Augur. The Augur team seems unconcerned about at least American regulation, with co-founder Joey Krug telling Reason Magazine that “Our friends in Washington, D.C. say the CFTC will probably just dismiss Augur and say it’s not a big deal.” However, it could play out differently, and Augur would still have to deal with regulators in other countries.
Gnosis is creating an open prediction market platform with transparent and equitable access to allow for accurate forecasts. Development on Gnosis, which runs on Ethereum’s blockchain, began in early 2015. The company is targeting their platform at the financial, insurance and information markets. Gnosis taking a hands-on approach and plans to develop applications for end users from their offices in Gibraltar.
Gnosis also has a plan to allow for integration with internet-enable devices that could provide data for the Gnosis platform. They foresee aggregating data produced by artificial intelligence systems to provide highly accurate forecasts. Effectively, this means there could be automated bots acting as oracles on the system.
Gnosis held a token sale in April 2017, raising $12.5 million USD for their GNO tokens. There is a fixed supply of 10 million GNO, of which 1.1 million are currently trading on exchanges. GNO began trading on exchanges at around $51 USD and is currently trading at $102 USD. GNO has a market cap of around $100 million USD.
The company holds GNO tokens itself and plans to use them to incentivize developers to work on the Gnosis platform.
The Gnosis platform has several layers. The Ethereum blockchain is the base, and on top of that is a core layer, a services layer and an app layer. The core layer is where the actual trading of outcome tokens happens. When a user wants to create a new event on Gnosis, she pays only the Ethereum gas costs. Participants who want to trade in the event then buy the outcome tokens associated with it. These participants pay a 0.5% fee when trading.
It seems that outcome tokens can be paid for with Ether. GNO tokens are used to generate another token called OWL via a smart contract. OWL tokens can then be used to pay the trading fees. OWL are intended to be a form of stablecoin, where one OWL can be used to pay for the equivalent of one USD in fees. However, fees can be paid with other ERC20 tokens. When this occurs, the token will be converted to GNO and removed permanently from circulation. However, Gnosis believes that OWL will most commonly be used to pay fees.
Gnosis provides examples of use cases in their whitepaper, such as determining the value of art, utilizing sensors from IoT networks or incentivizing the discovery of software bugs.
Gnosis plans to develop its own trading dApp. They also want developers to build on the platform, and plan to use GNO tokens to incentivize them to do this. How well they succeed in this should have a significant impact on the success of Gnosis.
Gnosis has been providing development updates on their blog. An alpha version called Gnosis Olympia has been running on Ethereum since 2017. Gnosis has been holding prediction tournaments on Olympia. They recently announced that a developer program to incentivize the building of applications on the platform, called Gnosis X, will begin on March 19. Developers will be able to win up to $100,000 in GNO tokens for creating applications in three different categories.
The full roadmap can be seen here.
Team and Advisors
The CEO and co-founder of Gnosis is Martin Köppelmann, who studied at the Hasso-Plattner Institute and was involved in the German Pirate Party. The CTO is Stefan George, who has a master’s Degree in Computer Science. The COO is Dr. Friederike Ernst, who holds a PhD from the Free University in Berlin and is general secretary of the German Blockchain Association, or Blockchain Bundesverband.
Gnosis shares some of its advisors with Augur, including Vitalik Buterin and Dr. Robin Hanson. Some other advisors to Gnosis are Joseph Lubin, co-founder of Ethereum and founder of ConsenSys; Jason Trost, founder and CEO of Smarkets; Jeremy Millar, chief of staff of ConsenSys; and James Slazas, a financial industry veteran.
There are 35 team members in total at Gnosis, though it is not clear how many are working solely on the prediction market platform.
When it comes to government regulation, Gnosis admits in their white paper that the platform may not be available in certain jurisdictions due to laws restricting gambling and betting sites.
Stox wants to bring decentralized prediction markets to mainstream audiences with their Ethereum-based platform. Stox was created by the team behind invest.com, an online financial services provider that was founded in 2014 and has 3 million users and $50 billion of revenue in 2016. invest.com allows users to trade in various financial products and is licensed to provide cross-border investment services in the European Union. Stox’s marketing strategy is to leverage invest.com’s user base to generate activity for Stox. The platform will run on invest.com’s IT infrastructure. Stox uses Bancor’s smart token protocol.
Stox held an ICO for its STX tokens in August of 2017. Floyd Mayweather participated in the promotion. According to TokenData.io, they raised around $33 million USD. There is a total supply of 57 million STX, with 42 million in circulation. It began trading at $1.21 USD per token and is currently at .29 USD.
50% of the tokens sold went to sale participants, 12.5% went to invest.com, Ltd., 10% went to the Stox team, and 27.5% went to Stox Ltd. to be used to bring strategic partners to the Stox ecosystem and as operational reserve.
Stox runs on Ethereum, and the STX token is an ERC20 token. In their white paper, Stox lays out a straightforward token model: STX is the only token necessary to do everything on the platform, including making trades, paying fees and providing bonds. Stox foresees allowing any event to be traded, such as election results, sports, entertainment or financial instruments. Users must stake STX when creating new events and fees are collected by event creators and market makers.
With regards to reporting outcomes, event creators can designate a centralized oracle, who is paid in STX. Disputes can be managed using consensus, where users essentially vote by staking STX, or by polling members about the outcome.
Stox is building an app for consumers. They also plan to work with incumbent providers of stock, options and currency trading and allow these providers to use the STX token in their systems as well. Stox are using Bancor as a token platform to handle liquidity issues.
· Released the beta version of the platform, called Sun Tzu
· Created a new scalar prediction type
· Translated their website into four other languages
· Hosted a product AMA
They are holding a Stox Cup IV tournament on the beta platform which runs through March. They are also hosting an “Ask Stox” series to allow people to ask them questions.
For their next steps, they plan to develop the Stox platform and app, launch with the invest.com customer network and create a software development kit so other companies can use the Stox app in their own systems.
Team and Advisors
Ophir Gertner is the Founder and Chief Strategist. He founded invest.com and has 15 years’ experience in financial markets. Some of the advisors include Eyal Hertzog, from the Foundation Council of Bancor, who is a technology entrepreneur and founder of video-sharing site MetaCafe; Uriel Peled, the CEO of CoinTree, an blockchain advisory firm; Itai Avneri, CEO of invest.com; and Joe Chen, CEO and founder of RenRen, a social networking service. There is a 20-person team.
Stox say in their white paper that governmental regulation and liquidity could be problems, but that they are committed to following all required rules. Since their plan appears to be more centralized than other prediction market platforms, they would likely find it more difficult than the other platforms to avoid online gambling laws.
Hivemind is an open-source protocol based on bitcoin that allows users to participate in prediction markets. It was created by Paul Sztorc and is based on an earlier proposal of his called Truthcoin. Sztorc began working on Truthcoin in 2014 and published a whitepaper. He later changed the name to Hivemind.
According to the website, Hivemind is in R&D stage with no ICO or token sale planned. No other financial information is publicly available.
Hivemind is a unique blockchain with two types of tokens: CashCoins (CSH) and VoteCoins (VTC). CSH are redeemable 1:1 for Bitcoin and represent value. Holders of CSH can create events and trade shares without owning VTC at all.
VTCs reflects a user’s reputation and can be compared in some ways to the REP coin in Augur. Outcomes of events are established by weighted-vote, according to users’ ownership of VTC. Owners of VTC also collect fees, which are paid in CSH, giving them an incentive to be accurate in their outcome results. If owners of VTC do not vote, then they lose their VTC.
Any user can create a prediction event in Hivemind by paying a fee, but there are only a limited number of decision slots available. Any holder of CSH can trade events, and event creators can earn fees from them. Consensus is achieved through proof-of-work and Bitcoin miners can mine the Hivemind blockchain for transaction fees.
Paul Sztorc is the founder of Hivemind. He is involved in Bitcoin development, is currently an economist at Bloq, a blockchain technical services firm, and formerly a researcher at Yale University in statistics. There are no other public team members.
With no apparent development work done in the past few years and no publicly known team beyond the founder, there remains a lot to do before Hivemind is up and running. However, the white paper is very detailed, and Mr. Sztorc is an experienced Bitcoin developer and appears to be very knowledgeable about prediction markets.
These proposed prediction market platforms are all vying to become the next big online gambling site. They claim to be superior to the previous generation due to the integration of blockchain technology, which would allow for global liquidity and improve the accuracy of predictions. However, there are still huge technological challenges to be overcome before they actually work. Further, if any of them gain popularity, it’s likely that governments around the world will take notice and attempt to regulate them. As a result, the odds are long that one or any of these platforms will be successful anytime soon.