Do Prediction Markets Need Web3?
Writers: Sid Rao, Mohit Pandit
The Lifecycle of a Bet
Investopedia says, “A prediction market is a market where people can trade contracts that pay based on the outcomes of unknown future events.” Essentially, a betting/wagering market. To understand betting markets a little better, let us break down the lifecycle of a bet:
Beliefs
At the belief stage, a prediction is just an opinion. When a person with an opinion puts their money where their mouth is, it becomes a bet, and are eligible to reap the rewards if an outcome favours their belief.
Beliefs are formed through a complex interplay of various cognitive, social, emotional, and environmental factors. Opinions can be generated by instantaneous beliefs or highly calculated thinking, and opinions are given more freely because there is no monetary loss to the person giving an opinion.
Bet
Bets are created when someone
- wants to capitalize on their belief
- Sees a very juicy outcome irrespective of a belief
The first kind of bet can come from a calculated opinion and the second one comes from a “bet small win big” attitude.
For any contract to be successful, there needs both a party and a counterparty:
- You betting $50 that Chelsea will win a match will require someone (or many) willing to put up a total of $50 that Chelsea will lose (assuming odds are 50/50)
- Margin trading on GMX where a maker opens a long position and a GLP is the counterparty
- Casino games like Roulette, Blackjack, etc. need a “House” to be a counterparty
Many a time there needs to be an incentive to be a counterparty because outcomes of an event aren’t equally likely. These incentives can come in many forms such as odds, the bonding curve in AMM, or even funding rates in perp/margin trading platforms.
Prediction market structures get more complicated in design as they focus on particular types of outcomes. Sportsbooks need special odds as almost none of the events are likely to have almost equal outcomes, and every major event (outcome of a League winner) might have many minor events that lead up to it (the outcomes of each league match).
There also needs to be correct execution of the contract. What if your counterparty refuses to pay up? This is why derivatives are essentially contracts that are enforced by law. On the blockchain, the contracts can be executed trustlessly depending on the outcomes.
So putting all the key events needed for a bet to take place, one needs:
- An event to occur (or not occur), and post the event/game contract
- Ensure that enough parties have opinions on these events (maker demand)
- Ensure there are counterparties to these parties (taker demand)
- Ensure settlement
- Ensure there is no manipulation
Outcomes
“Gambling games promote an ‘illusion of control’: the belief that the gambler can exert skill over an outcome that is actually defined by chance.” — Dr. Luke Clark
Outcomes are the closure of a bet on the event. Once the outcome is decided the wager is completed.
Do prediction markets need Web3?
Let us look at each of the above-mentioned criteria to create a betting market and try to make a case for web3:
Event/Game Creation
There is no clear blockchain use case for this except permissionless posting of an event. Permissionless posting is a bug, and not a feature because this can create high redundancy for the same event and make the bettor experience bad. Wager creation can be based on an event, or also create a game like on-chain roulette, or blackjack.
The event could also be Price discovery. We see prediction markets on Aevo, for tokens that have not been launched yet. This gives a good indicator of what the market is thinking about the price of a token.
Parcl is also creating a prediction market for better price discovery in Real Estate. It gives a homeowner a ballpark figure of what their house is worth, and a buyer looking to purchase real estate in a city a ballpark budget.
The price discovery use case is also a function of liquidity in the event contract which is why the next section is important
Maker Demand
This cannot be controlled by the blockchain because this is completely driven by off-chain actions like marketing or games that are built into a product.
Companies targeting price discovery must focus hard on generating as much maker volume to get the most accurate price for a particular asset.
Counterparties
Now this is where it gets interesting. Counterparties can be incentivized with juicy odds to take a gamble where the event outcome is potentially a sure shot. In the image below you can see that there is a potential to make $200 with a $0.50 bet because of the sheer mismatch in the Polymarket order book.
The other way to do it is like Augur Turbo, where each market is an isolated market that works on Balancer AMMs. Here the LPs behave as sort of the counterparties for different markets. While this is a great structure to avoid heavy reliance on odds calculation (or acquisition), it makes it a bad experience for posting prediction events.
For price discovery order books like Aevo, if there is no liquidity, there will be times when the platform itself will have to behave as the counterparty. This is not ideal. Especially in a market where the floor is unknown.
The other way to do it is to create a counterparty LP pool that behaves like “The House”. Very much like how Azuro and WINR are doing. There is a liquidity pool that will act as a counterparty to the bettors. Parcl has a USDC liquidity pool to act as a counterparty to traders longing for or shorting Real Estate prices in different cities.
Both these protocols have proven to do well:
Models like these show some good product market fit, where frontends only need to worry about bettors placing bets on the platform without the requirement of managing order books or making the tradeoffs that an AMM brings.
You can think of these models as a Uniswap v4 with different frontends using the underlying liquidity (which are analogous to hooks).
WINR protocol has 1 casino betting frontend and another margin trading protocol that offers up to 1000x leverage which ensures high pool utilization but could be very dangerous to the pool.
Ensure settlement
As soon as the event is completed, the wager needs to be settled.
- In an AMM structure, everything is on-chain and is settled on the contract.
- With the Polymarket model of order books, the order book is maintained off-chain. Polymarket can block withdrawals if needed
- With Azuro frontends like Bookmaker.xyz, there is no need to make deposits. Every bet placed is treated as an individual transaction. The only off-chain components are the calculation of odds, and also the data feeds.
Ensure there is no manipulation
If there is a centralized data provider, and the data feed is manipulated by the provider, this can adversely affect the outcome for a maker/taker. This is one of the main reasons why most Web3 prediction markets use Oracle systems like Chainlink. Using an oracle will have the trade-off of latency vs. data integrity. Even while choosing oracles, platforms can choose between first and third-party oracles with tradeoffs on latency. Latency is most important in more high-paced events.
When it comes to playing casino games, the randomness function cannot be compromised depending on where it comes from.
Chainlink and other oracles like Supra and Pyth try to mitigate as much manipulation via aggregation, but provenance is still always going to be an issue in a vast market
Successes and Failures of Existing Applications
When we look at Crypto x Prediction markets, the most success we have seen is with crypto being used as an asset to place wagers on applications like Stake.com and Rollbit.
While apps like Polymarket have seen sparse success depending on certain macro events, it is not a platform where there can be consistent volumes as there is a huge gap in the environment of an event and the platform.
Crypto x Prediction Markets PMF has somewhat been found in “House” pool systems like Azuro and WINR. There is a clear use case where newer frontends who want to focus on a particular type of prediction market need to focus only on the demand side. They can tap into systems like Azuro and WINR which in turn provide holders of stablecoins top-in-class yields (40–60% APY at the rate at which they are going).
Regulation on betting apps and online casinos is very tight in most countries. Protocols like Azuro and WINR also might face lower regulatory pressure than someone like Rollbit.
Crypto only allows as much participation as the front end does. There are no completely permissionless and trustless crypto prediction markets
What we are excited to see are the possible successes of an application like Parcl which brings transparency to quite an illiquid asset class. By first principles, it looks like it has the right structure to succeed in its goal of price discovery.
The top use cases of Web3, are the counterparty pool structure that allows for different types of prediction markets to be built, and then hopefully the success of prediction markets for better price discovery as a use-case.
The higher the crypto market cap goes, and the more people have disposable capital on-chain, the prediction market industry could be very lucrative and/or useful.
