Original title: Prediction Markets: Bottlenecks and the Next Major Unlocks
Original author: Mikey 0x, 1kx
Compiled by: Elvin, ChainCatcher
abstract
1. How prediction markets work
2. Bottlenecks that hinder wider adoption of prediction markets
Supply Side
Demand side
3. Solution
Supply Side
Demand side
4. Other ways to increase adoption
Prediction Market: Development Bottlenecks and the Next Key Opportunities
Augur, the first on-chain prediction market, was one of the first applications to be launched on Ethereum. Its vision was to allow anyone to bet any size on anything. Augurâs vision was not realized many years ago due to numerous problems. Lack of users, poor settlement UX, and high gas fees led to the closure of the product. However, we have come a long way since then: block space is cheaper and order book design is more efficient. Recent innovations have solidified the permissionless and open source nature of cryptocurrencies, allowing anyone to become a participant in the global liquidity layer by providing liquidity, creating markets, or placing bets.
Polymarket has become the market leader with around $900 million in volume to date, while SX Bet has accumulated $475 million to date. Still, there is a lot of room for growth compared to the massive size of the traditional prediction market subcategory of sports betting. In the United States alone, sportsbooks process over $119 billion in volume in 2023. This number stands out even more if all other countriesâ offline and online sports betting volumes as well as other types of prediction markets such as politics and entertainment are taken into account.
This article aims to break down how prediction markets work, the current bottlenecks that need to be addressed, and some of the ways we think these issues can be addressed.
How do prediction markets work?
There are several ways to design prediction markets, most of which can be divided into two categories: order book models and centralized AMM models. Our view is that order book models are the superior design choice because they allow better price discovery, achieve maximum composability, and ultimately lead to scaled trading volume.
Order Book Model
For the order book model, each market has only two possible predefined outcomes: yes (Y) and no (N). Users trade these outcomes in the form of shares. At market settlement, the correct shares are worth $1 and the incorrect shares are worth $0. Before the market settles, the shares may trade between $0 and $1.
In order for share trading to occur, liquidity providers (LPs) must exist; in other words, they must provide buy and sell orders (quotes). These LPs are also known as market makers. Market makers provide liquidity in exchange for a small profit on the spread.
Letâs take a specific market as an example: if something is equally likely to happen, like a coin toss coming up heads, then in theory both the âYesâ and âNoâ shares should trade at $0.50. However, like any financial market, there is usually a spread and therefore slippage. If I want to buy the âYesâ shares, my trade price may end up being closer to $0.55. This is because my counterparty, a market maker, is intentionally overestimating the true odds to make a potential profit. The counterparty may also sell the âNoâ shares for $0.55. The $0.05 spread on each side is the market makerâs compensation for providing the quote. The spread is driven by implied volatility (the expectation of price movement). Prediction markets inherently have a guaranteed realized volatility (actual price movement), simply due to the design that the shares must eventually reach $1 or $0 on some predetermined date.
Market maker scenario example:
The market maker sells 1 share of âYesâ at $0.55 (which is equivalent to buying 1 share of âNoâ at $0.45)
The market maker sells 1 share of âNoâ at $0.55 (which is equivalent to buying 1 share of âYesâ at $0.45)
The market maker now has 1 No share and 1 Yes share, paying a total of $0.90
Regardless of whether the coin lands on heads or tails, the market maker will exchange $1 and earn a $0.10 spread
The other main way prediction markets are settled is through centralized AMMs, which are used by Azuro and Overtime. I wonât go into these models in this article, but the analogy in DeFi is GMX v2. Capital is centralized, and as the only counterparty to platform traders, the pool relies on external oracles to set prices for users.
What is the bottleneck of the current prediction market?
Prediction market platforms have been around and discussed long enough that if there was true product-market fit, escape velocity would have already occurred. The current bottleneck can simply be attributed to a lack of interest on both the supply (liquidity providers) and demand (bettors) sides.
Supply-side issues include:
1. Lack of liquidity due to volatility: Polymarketâs most popular markets tend to be conceptually novel markets that lack relevant historical data, making outcomes difficult to predict and accurately price. For example, predicting whether a CEO, such as Sam Altman, will âreturn to his position after rumors of potential AGI mishandlingâ is difficult because there are no past events that closely resemble this situation. Market makers will set wider spreads and less liquidity in uncertain markets to compensate for implied volatility (i.e., the price of the Sam Altman CEO market has moved wildly, with consensus flipping 3 times in less than 4 days). This makes large investors who want to place large bets less interested.
2. Lack of liquidity due to lack of subject matter experts: Although hundreds of market makers are rewarded every day on Polymarket, many long-tail markets lack liquidity due to a lack of participants with expertise. For example, markets such as "Will celebrity X be arrested or charged for X?" or "When will celebrity X tweet?" This situation will change over time as more types of prediction markets are introduced, data becomes richer, and market makers become more professional.
3. Information asymmetry: Since the buy and sell quotes provided by market makers can be traded by any taker at any time, the latter has the advantage of making positive expected value bets when obtaining favorable information. In the DeFi market, these types of takers can be called harmful flows. Arbitrageurs on Uniswap are good examples of harmful takers because they use their information advantage to continuously extract profits from liquidity providers.
In one Polymarket market, âWill Tesla announce a purchase of Bitcoin before March 1, 2021?â, one user bought $60,000 worth of âYesâ shares at odds of about 33%. This market was the only one this user had ever participated in, and it can be assumed that this user had favorable information. Leaving aside the legality issue, the market maker providing the quote had no way of knowing at the time that the taker/bettor had this favorable information, and even if the market maker initially set the odds at 95%, the taker might still have bet because the true odds were 99.9%. This resulted in the market maker facing a certain loss. In prediction markets, it is difficult to predict when harmful traffic will occur and how large it will be, making it even more difficult to provide tight spreads and deep liquidity. Market makers need to price in the risk of harmful traffic that could occur at any time.
The main issues on the demand side are:
1. Lack of leverage tools: Without leverage tools, prediction markets are relatively unattractive to retail investors relative to other crypto speculative tools. Retail investors want to create âgenerational wealth,â which is more likely to be achieved on memecoins than on capped prediction markets. For example, early bets on $BODEN and $TRUMP brought more upside than the âyesâ shares on the prediction market for Biden or Trump winning the presidential election.
2. Lack of exciting short-term markets: Retail bettors have no interest in placing bets that will be settled months later, a conclusion that can be demonstrated in the world of sports betting, where much of the retail trading volume now occurs on live betting (ultra-short term) and daily events (short term). There are not enough short-term markets to attract a mainstream audience, at least not yet.
What are the solutions to these problems? How can we increase transaction volume?
On the supply side, the first two issues, related to illiquidity due to volatility and illiquidity due to lack of expertise, will naturally decrease over time. As trading volume on the various prediction markets grows, the number of professional market makers and those with higher risk tolerance and capital will also grow.
However, rather than waiting for these issues to abate over time, the illiquidity problem is addressed head-on through a liquidity coordination mechanism originally invented in the DeFi derivatives space. The idea is to allow passive stablecoin depositors to earn yield through a vault that deploys market-making strategies in different markets. This vault will act as the primary counterparty to traders. GMX was the first protocol to do this through a pooled liquidity provision strategy that relies on oracles for pricing, while Hyperliquid was the second notable protocol to deploy a native vault strategy, but with the distinction that liquidity is provided on CLOBs. Both vaults have been profitable over time because they are able to act as counterparties to the majority of harmless flows (retail users who tend to lose money over time).
Hyperliquid's vault PNL has been growing over time
Native vaults make it easy for protocols to bootstrap liquidity on their own, without having to rely on others. They also make long-tail markets more attractive; one reason Hyperliquid has been so successful is that newly listed perpetual assets include a lot of liquidity from day one.
The challenge of building a treasury product for prediction markets is preventing harmful flows. GMX prevents this by attaching high fees to its trades. Hyperliquid uses a market maker strategy with wide spreads, a 2 block delay on taker orders to give market makers time to adjust their quotes, and prioritizes market maker order cancellations within a block. Both protocols create an environment where harmful flows do not enter because they can find better price execution elsewhere. In prediction markets, harmful flows can be prevented by providing deep liquidity with wide spreads, selectively providing liquidity to markets that are not susceptible to information advantages, or hiring savvy strategists with information advantages.
In practice, a native vault could deploy an additional $250,000 in liquidity, bidding at 53 cents and asking at 56 cents. Wider spreads help increase potential vault profits because users accept worse odds when placing bets. This is different from setting bids at 54 cents and 55 cents, where the counterparties may be arbitrageurs or savvy traders looking for a good price. This market is relatively less susceptible to information asymmetry issues (less inside information and insights are often disclosed to the public quickly), so the expectation of harmful flows is lower. Vaults can also use information oracles that provide insights into future line changes, such as pulling odds data from other betting exchanges or collecting information from top political analysts on social media.
The result is a deeper level of liquidity for bettors, who are now able to place larger bets with less slippage.
There are several ways to solve or at least reduce the information asymmetry problem. The first ones are about order book design:
1. Gradual Limit Order Book (GLOB): One way to combat harmful flow is to increase the price by combining the speed and size of orders. If buyers are certain that an event will occur, the logical strategy would be to buy as many shares as possible at a price below $1. In addition, if the market eventually gets favorable information, it is also wise to buy quickly.
Contro is implementing this GLOB philosophy and is launching as a cross-rollup on Initia.
If the Tesla $BTC market happened on the GLOB model, the taker would have to pay more than 33% of the âyesâ shares because of the âslippageâ that occurs given the speed at which the orders are combined (a slice) and the size (huge). He would still make a profit knowing that the âyesâ shares will eventually appreciate to $1, but this would at least cover the market makerâs losses.
One could argue that if the taker is just long on a DCA strategy, they could still afford very little slippage and pay closer to 33% per âyesâ share, but in that case at least give the market maker some time to withdraw its quote from the book. Market makers may withdraw for several reasons:
Suspected harmful traffic because such a large taker order came in
It was certain there was harmful traffic because it checked the recipients' profiles and found they had never placed a bet before.
wants to rebalance its inventory, and no longer wants to be too one-sided due to the number of âyesâ shares it is selling and the ânoâ shares it is accumulating as a result â perhaps the market maker initially has $50k worth of orders on the ask side at 33% odds and $50k worth of orders on the bid side at 27% odds â its initial goal is not directional bias but neutrality in order to earn profits through symmetrical liquidity provision
2. Winner Take: There are many markets that redistribute part of the profits to those with favorable information. The first example is in peer-to-peer web2 sportsbooks, notably Betfair, where a fixed percentage of the user's net winnings is redistributed back to the company. Betfair's take really depends on the market itself; on Polymarket, it may be reasonable to charge a higher net winning take for more novel or long-tail markets.
This concept of redistribution exists in DeFi in the form of order flow auctions. A back-running bot captures value from information asymmetry (arbitrage) and is forced to give it back to the person participating in the transaction, which could be a liquidity provider or the user making the trade. Order flow auctions have seen a lot of PMFs to date, and CowSwap* is pioneering this category with MEVBlocker.
3. Static or dynamic taker fees: Polymarket currently has no taker fees. If this were implemented, proceeds could be used to reward liquidity provision in high volatility markets or markets that are more susceptible to harmful flows. Alternatively, higher taker fees could be set on long-tail markets.
On the demand side, the best way to address the lack of upside is to create a mechanism that allows it. In sports betting, parlays have become increasingly popular with retail bettors because they offer the chance to "win big." A parlay is a bet that combines multiple individual bets into a single bet. To win the parlay, all individual bets must win.
User wins over $500,000 with initial bet of $26
In cryptocurrency-native prediction markets, there are three main ways to increase upside for users:
Parlays
Perpetuals
Tokenized leverage
Parlays: Technically, it is not feasible to implement this on Polymarketâs books, as bets require upfront capital and the counterparty is different for each market. In practice, a new protocol could fetch the odds from Polymarket at any given time, price the parlays, and act as the single counterparty for the parlays.
For example, a user wants to bet $10 on the following:
These bets have limited upside if placed individually, but when combined into a chain, the implied return soars to about 1:650,000, meaning that if every bet is correct, the bettor can win $6.5 million. Itâs not hard to imagine how chaining can gain PMF (product-market fit) among crypto users:
The cost of participation is low, you can win a lot with a small investment
Sharing a jackpot ticket can go viral on Crypto Twitter, especially if someone wins a big prize, which creates a feedback loop with the product itself.
Supporting parlays brings challenges, namely counterparty risk (what happens when multiple bettors win large parlays at the same time) and odds accuracy (you donât want to offer bets where you underestimate the true odds). Casinos have solved the challenge of offering parlays in the sports world, and it has become the most profitable part of sports betting. The profit margins are about 5-8 times higher than offering single market bets, even if some bettors get lucky and win big. Another added benefit of parlays is that there is relatively less harmful traffic compared to single markets. The analogy here is: why would a professional player who lives on expected value put money into the lottery?
SX Bet, a web3 sports betting application chain, launched the world's first peer-to-peer syndicated betting system and achieved $1 million in syndicated trading volume in the past month. When a bettor "requests a syndicate", SX creates a private virtual order book for the syndicate. Programmatic market makers listening through the API will then have 1 second to provide liquidity.
Perpetual Prediction Markets: This concept was briefly explored in 2020 when leading exchange FTX offered a perpetual contract for the US election results. You could go long $?????? at a price that would redeem $1 per share if he won the US election. As his actual odds of winning changed, FTX had to change margin requirements. Creating perpetuals for markets as volatile as prediction markets presents a lot of challenges for margin requirements as the price can go from $0.90 to $0.10 in a second. As a result, there may not be enough collateral to cover the losses of someone who went long in the wrong direction. Some of the order book designs explored above can help compensate for the fact that prices change quickly. Another interesting thing about the FTX $TRUMP markets is that we can reasonably assume that Alameda is the primary market maker for these markets, and without locally deployed liquidity, the order book is too thin for a large number of trades. This highlights the value of local liquidity vault mechanisms for prediction market protocols.
LEVR Bet and SX Bet are currently developing a perpetual sports betting market. One advantage of sports betting is that the price fluctuations of "yes" or "no" shares will be smaller, at least most of the time. For example, a player shooting a field goal may increase the team's chance of winning the game from 50% to 52%, because on average a team may shoot 50 times per game. A 2% improvement on any given shot is manageable from a liquidation and closeout requirement perspective. Offering perpetual contracts at the end of a game is another matter, because someone may hit the "game-winning shot" and the odds may flip from 1% to 99% in half a millisecond. One possible solution is to only allow leveraged betting to a certain extent, because any event after that may change the odds too much. The feasibility of perpetual sports betting also depends on the sport itself; a hockey goal is more likely to change the expected outcome of a game than a basketball shot.
Tokenized Leverage: A lending market that allows users to borrow against their prediction market positions, especially those that are long-term, could increase trading volume among professional traders. This could also lead to more liquidity, as market makers can borrow against positions in one market in order to make markets in another. Tokenized leverage may not be an interesting product for retail bettors unless there is an abstracted loop product like the ones that Eigenlayer is gaining traction with. The overall market may still be too immature for such an abstraction layer to exist, but these types of loop products will eventually appear.
Besides the pure supply and demand side of things, there are other small ways to increase adoption:
From a UX perspective: Switching the settlement currency from USDC to a yield-earning stablecoin will increase participation, especially in long-tail markets. This has been discussed a few times on Twitter; holding a market position that expires at the end of the year has a significant opportunity cost (e.g., earning 0.24% APY on a bet on Kanye West to win the presidential election, rather than earning 8% APY on AAVE).
Furthermore, increased gamification aimed at improving retention can really help attract more users in the long run. In the sports betting industry, simple things like âdaily betting streaksâ or âdaily contestsâ have worked well.
Several industry-level tailwinds will also increase adoption in the near future: the growth of virtual and on-chain environments will unlock a whole new level of speculative demand, as the number of short-term events will eventually be infinite (think AI/computer-simulated sports) and the level of data will be abundant (making it easier for market makers to price outcomes). Other interesting crypto-native categories include AI games, on-chain games, and on-chain data in general.
Accessible data will lead to an increase in non-human gambling activity, more specifically, an increase in betting activity by autonomous agents. Omen is pioneering the idea of ââAI agent bettors on Gnosis Chain. Since prediction markets are a game where the outcomes are defined, autonomous agents will become increasingly skilled at calculating expected value, potentially with far greater accuracy than humans. This reflects the idea that AI may have a hard time predicting which meme coins will take off because there is a more âemotionalâ element to what makes them successful, and humans are currently better at feeling emotion than AI.
All in all, prediction markets are a fascinating user product and design space. Over time, the vision of allowing anyone to bet on anything at any scale will become a reality. If you are building something in this space, whether it is a brand new protocol, a liquidity coordination platform, or a new leverage mechanism, please get in touch. I am an avid user and am more than happy to give feedback.
Thanks to Peter Pan, Shayne Coplan, Sanat Kapur, Andrew Young, taetaehoho, Diana Biggs, Abigail Carlson, Daniel Sekopta, Ryan Clark, Josh Solesbury, Watcher, Jamie Wallace, and Rares Florea for their feedback and reviewing this post!
Disclaimer:
*Denotes 1kx portfolio investment.