Augur vs. Gnosis and the Race to Decentralize Vegas: Part 1
This post will be the first in a series about decentralized prediction markets. In part 1, I will outline what a prediction market is and why it benefits from blockchain technology. In part 2 and 3, I will go into the details of Augur and Gnosis, two projects leading the charge to make this idea a reality. I will conclude with Part 4, where I will compare both projects.
A prediction market is a market that allows people to trade on the outcomes of events. In layman’s terms, it is a betting platform in which anything can be bet on. It is called a “prediction” market because by utilizing the knowledge of crowds we can discern more accurate predictions than through traditional means. For example, the incumbent prediction market, Vegas, has a proven track record of accurately predicting sporting outcomes. The reason for this is that if someone stands to lose or gain money on the outcome of the event, they will take significantly more care in their prediction to make sure they’re correct.
A prediction market also takes into account non-public information. If you know something no one else knows, then you are going to make a strong bet on a certain outcome. This is information that would never come out through a forecast, survey, or poll, but will be considered in a prediction market. For this reason, a prediction market improves becomes more accurate as more people participate.
The best example of a well-functioning prediction market is sports betting. A lot of participants are involved in the outcomes of sports events, worth hundreds of billions of dollars. Financial markets are also a type of prediction market (though not in the purest sense) because a lot of people place big bets on the outcomes of companies. For that reason, we consider the financial markets and the Vegas odds to be fairly efficient.
How it Works
A prediction market works by putting all outcomes on a scale of 0% to 100%, with the price of “shares” adjusting based on new bets that are placed. For example, if an outcome has a 50% chance of happening, you would buy shares of that outcome for $0.50. If that outcome was correct, then your shares become worth $1 each.
Let’s run through a simplified, high level example of a prediction market for an election in which 3 candidates are running: Alice, Bob and Carol.
The total amount of dollars placed on each candidate determines the odds and share price for each candidate ($60, $35, and $5, in this case). Now let’s say that I want to make a bet on Bob, because I believe he has a better than 35% chance of winning. I decide to spend $40 giving me around 114 “shares” (40 / 0.35), with the hope that if Bob wins, my shares will become worth a total of $114. But if he loses I walk away with nothing.
However, my $40 bet on Bob changes the odds for anyone else wanting to bet (the total dollars placed on Bob increased from $35 to $40). The market has taken into account my prediction and now gives Bob a greater chance of winning, in-turn increasing the purchase price of his shares.
While this example is extreme due to the low amount of total dollars bet, it still shows how the mechanics of a prediction market work. Once a sufficient number of people (more specifically, capital) have placed their bets, we reach odds that should be more or less accurate. And if they’re not, people with good information would bet until the odds reach equilibrium. (Note that in reality, a prediction market will have a bid ask spread).
Applying it to the Blockchain
Blockchains are terribly slow, wildly inefficient, and massively complicated. This makes them not suitable for the majority of use cases. However, the purpose of a public blockchain is to remove the need for a trusted third party, making the application borderless, permissionless, and censorship-resistance. When evaluating a blockchain project it is important that it benefits from these three points.
- Borderless: Is it important that the app can be accessed by anyone irrespective of where they live?
- Permissionless: Is it important that the app can’t prevent anyone from using it?
- Censorship-Resistant: Is it important that what goes on in the application cannot be fixed or gamed by a central party?
If the answer to any of these questions is no, then a blockchain is not required, and a centralized system would work better.
So why does a prediction market work better when it’s borderless, permissionless, and censorship-resistant? Quite simply, the more people who are involved in the market, the more accurate it will be. The odds for a basketball game will be significantly more accurate if they consider the bets of 10,000 people, as opposed to 10 (keeping capital per person constant). As shown in my election example, a low amount of total dollars bet can lead to extreme variations in the odds, significantly reducing its accuracy.
By moving a prediction market to the blockchain you can open the platform to the entire world, and in turn account access global information towards the odds of a given outcome. However, most people today do not have easy access to a prediction market. There are some betting websites, but laws vary significantly by jurisdiction which complicates things for the providing company.
Another important function of removing a centralized third party from a prediction market is that anyone can make an event. This is something that should not be taken lightly. The incumbent prediction markets might have large sporting or political events, but if you want to bet that it will rain tomorrow in San Francisco, you’re out of luck. With a decentralized platform, a market could be created on anything, as long as you have at least one person on the other side. Events like this are not just unprofitable for centralized entities but they also suffer from the local knowledge problem when knowing which markets to make.
Perhaps the most important reason to move prediction markets to the blockchain is for regulatory reasons. Despite being a “prediction market”, it is essentially gambling. And as we know from current gambling or prediction market websites, regulatory compliance is perhaps the greatest hindrance. By moving to a sovereign-grade resistant blockchain such as Ethereum, we can expect a prediction market to fare sufficiently better against jurisdictional regulations.
Use Cases for Prediction Markets
A massive, global, permissionless prediction market will create use cases that currently are not possible today. While there will be others, there are three primary ways I could see this innovation being used: recreational betting, predictions, and hedging.
I define recreational betting as someone betting on something in which the outcome would not directly affect them (financially) had they not bet on it. Examples of this would be: sports, elections, random public events, etc. An open system would allow this to grow beyond politics and sports to any event that two sides want to bet on, greatly increasing total market size.
I define the prediction use cases as one in which someone who wants to get the most accurate probability of an outcome occurring. If you have a decision to make that is dependent on an outcome of an event, but you have no clue as to the probability of that outcome, you could look to see what the prediction market says the odds are, given it has a liquid enough market.
A great example of this would be the 2016 US presidential election, where the probabilities of each candidate’s success were wildly wrong, due to the inaccuracy of surveys and polls. If you ran a business dependent on the policies of Hillary, you likely made some poor decisions in hindsight leading up to the election. Had there been a decentralized prediction market, the odds would most likely have been much closer, making you think twice about your decision.
Another interesting way this could be used would be for current predictions where innovation has stifled due to lack of incentive. If decentralized prediction markets were to go mainstream, I could see them replacing and improving things like weather forecasting. With significant financial incentive to improve weather predictions, the entrepreneurial meteorologist would innovate in ways she traditionally had no motive for.
The use case that I see as the biggest wildcard in its potential impact is hedging. Hedging is the ability to reduce one’s risk exposure to a given event. This is already very common in the financial world through derivatives such as forwards and futures. For example, an apple farmer can lock in the price he will sell his apples for in the future by making a deal with the apple pie maker. This financial transaction is called a forward contract and removes the risk for both parties from volatile apple prices.
However, today it isn’t possible to hedge risk in the more random or one-off events that can exist, especially in less developed countries. Let’s look at a simplified example of a Brazilian farmer who needs a certain level of rainfall to grow his crops, to understand the power of hedging.
Our farmer needs at least 30 inches of rainfall in order to have a successful harvest. If there is < 30in of rainfall the farmer will not make any money, but any amount greater than that and he can have a full harvest netting him $10,000. The farmer is nervous because experts are predicting lower rainfall than usual, putting the farmer in danger of not being able to make any money this year. However, he finds (or creates) a prediction market for the odds of Brazilian rainfall being < 30in, allowing him to hedge the downside risk of lack of rainfall.
If the prediction market has the probability of less than 30 inches of rainfall as 40%, the farmer can buy $1000 worth of “shares” at $0.40 each. He would then set to make $2500 if there is less than 30 inches of rain. This money will allow him to feed his family and stay out of debt in the case he isn’t able to reap a harvest.
While I understand this scenario is completely hypothetical, you can still see the value in giving someone the ability to hedge their risk in a way that was previously not possible (or significantly more difficult). It’s impossible to predict all the ways that this form of hedging would be used, but the potential value is there especially in lesser developed countries.
The value of blockchains for hedging risk like this opens the possibility for anyone to hedge as long as someone’s on the other side. In extension of this I can see all sorts of new types of insurance being created.
When an event occurs, it is imperative that it matches the truth of what actually happened. This is not an issue in traditional prediction markets because the law protects those who bet correctly to receive their money. A casino can’t payout those who bet on the Cavaliers if the Warriors were the winning team. Since blockchains are not controlled by a single entity, they need have specific measures to ensure that a certain outcome is the correct one. Additionally, blockchains are immutable, making decisions final in most cases, which further exacerbates the importance of this issue. The solution is something called an oracle.
An oracle is simply a way to pull non-blockchain information into the blockchain, which can be achieved in a few different ways. One would be to configure an API, such as programming the contract to pull the results of a sports game from ESPN.com. However, this wouldn’t be terribly secure since there is the potential for a rogue employee to switch the results at the time the oracle was going to pull the data. A better oracle could pull the results of the game from multiple sources (ESPN, FOX Sports, etc), which would further reduce the likelihood of pulling wrong information.
An oracle could also rely on humans (token holders) to determine the outcome through a vote. Of course, the more parties you have the more secure this method would be. But compared to the API oracle, it would require more coordination and incentivization for the parties involved.
Oracles are extremely important because if the information on the blockchain doesn’t reflect reality, you lose the integrity of the platform. Different decentralized prediction market applications deal with this in different ways, which I will explore extensively in future posts.
While prediction markets currently exist in specific jurisdictions for specific topics, having a global platform accessible to anyone is what I believe to be one of the most profound blockchain use cases.
There are currently two projects that are attempting to create a decentralized prediction market: Augur and Gnosis. In Part 2 of this post I will go into detail on the inner workings of Augur, followed by Gnosis in Part 3, and a comparison in Part 4.
Disclaimer: The views expressed in this article are solely the author and do not represent the opinions of the author on whether to to buy, sell or hold shares of a particular cryptocurrency, cryptographic asset, stock or other investment vehicle. Individuals should understand the risks of trading and investing and consider consulting with a professional. Investors should conduct their own research independent of this article before purchasing any assets. Past performance is no guarantee of future price appreciation.