Decentralized Prediction Markets — Past to Present
Why recent developments in DeFi infrastructure have set the stage for decentralized prediction markets to take off.
By Ariana Fariab, Kelsey Lawler and Kyle DiPeppe
To say prediction markets were early to the DeFi ecosystem would be an understatement. In the 2010s, if you were building a DApp (Decentralized Application) you had one viable option: Ethereum. With their alpha version live in mid-2015, the first decentralized prediction market (DPM), Augur, was also one of the first DApps on Ethereum, placing DPMs among the first structures to try decentralization.
Despite Augur’s initial traction and amassing media attention, public excitement, and support from big names in crypto, the platform’s struggle for long-term success came down to user experience. Slow, unreliable, unscalable infrastructure rendered a poor experience, stunting adoption. Though blockchain infrastructure offerings have improved over the years, it wasn’t until recently that the space underwent the shifts necessary for DPM platforms to provide an optimal user experience. This piece explores why DPMs may have been too early for their time, and why we believe the timing is right now.
. . .
A Brief Primer on Prediction Markets
Prediction markets (PMs) enable participants to trade on the outcomes of future real-world events. Markets can pose questions such as ‘Will candidate X win the presidential election?’ and participants take positions tied to possible outcomes. Leading up to the underlying event, PM share prices are dictated by the wisdom of crowds. These prices draw on vast reserves of crowd knowledge and reflect the market participants’ collective consensus on the probability of each outcome occurring. Once the actual outcome is known, share prices are settled such that shares attached to the prevailing outcome increase in value and all others resolve to zero. Prediction markets can cover any event with objectively verifiable results, such as political events, sporting events, crypto prices, etc.
The ideas that shaped prediction markets date back to the twentieth century, with economist Friedrich Hayek’s essay The Use of Knowledge in Society (1945), which argued that prices aggregate and reflect information held by a group. A paper on long run prediction market accuracy published in the International Journal of Forecasting (2008) measured the forecasting performance of the Iowa Electronic Markets (IEM), one of the earliest prediction market platforms, against political polls; when compared to 964 polls on presidential elections spanning 1988–2004, the market was closer to the real outcome 74% of the time. Following the IEM, a handful of prediction market platforms emerged over the years, including The Hollywood Stock Exchange (one of the first online platforms), InTrade, PredictIt, and, more recently, Kalshi. While the IEM focused on political events, these later platforms expanded to include the variety of topics we see today.
. . .
Decentralized Prediction Markets Emerge
Born to combat what its founders called “the three Cs of centralization” (closed, constrained, costly), the first DPM platform, Augur, was founded by the Forecast Foundation in 2014 and launched on Ethereum mainnet in July 2018, after three years of development and testing. The project’s founders believed factors such as capital controls, regulations, investment caps, and high platform fees limited the utility and forecasting powers of centralized prediction markets. Augur also tackled issues inherent to centralized markets, centralized: order books, market creation, and resolution sources.
In the days following its launch, Augur saw hundreds of users trade nearly $1.5M on the platform. Just a day post-launch, it was already a top-ranking DApp by user numbers, placing #1 in its category and #5 overall. Leading up to Augur’s launch, the prospect of decentralized prediction markets stirred up a whirlwind of excitement from both prominent figures in crypto and native and non-native would-be users. The narrative surrounding DPMs was bullish. As Multicoin Capital Co-founder Kyle Samani put it to CoinDesk a month before the launch, “I think Augur’s going to be probably the most widely used DApp when it launches.”
Problems Arise
But the platform’s early success was short-lived. User numbers began declining rapidly, cascading from a peak of 265 on launch day to 113 just 6 days later, and a mere 37 on August 8th (~ month after launch). What caused this mass exodus from the highly anticipated platform? It all came down to user experience. As Augur co-founder Joey Krug put it himself, the platform was slow, expensive, and hard to use. Frustrated, early users took to Reddit and Discord declaring Augur unusable. The most common complaints were crippling slowness, high gas fees, volatility issues associated with markets being priced in ETH, inconsistent blockchain data, and low liquidity. Most of these usability issues (slowness, gas fees, inconsistencies) stemmed from problems in the DApp’s underlying infrastructure. Part of the reported slowness was the time it took just to gain access. Augur initially utilized a free third-party RPC node, which users reported taking hours to sync with the Ethereum blockchain. While this node was responsible for the appalling sync times and frequent inconsistent data, the root of slow transaction speeds and high gas fees was Ethereum.
As is commonly known, Ethereum suffers from scalability issues. With a throughput of 15–45 transactions per second (TPS), it comes as no surprise that transactions could take anywhere between 15 seconds to 6+ minutes to reach finality, an issue only exacerbated as the network became more popular. Network congestion and low throughput were also responsible for the high gas fees. As more users flooded the Ethereum network, the less usable it became. Simply put, Augur’s underlying infrastructure was not equipped to support it. Although the application itself was working, the infrastructure powering it was not. As a HackerNoon post put it, Augur had “built a house in an area without electricity, water, communications, or roads”. Even prior to launch, the project’s founders foresaw difficulties stemming from the underlying blockchain. In the Augur Master Plan, Krug asserted that “probably the hardest and furthest off” piece of the puzzle would be enabling more TPS, which he insightfully predicted would require implementing PoS, sharding, and off-chain scaling solutions. In another post, published after launch, Krug espoused that speed issues would be solved by Layer-1 scaling; things like state channels and plasma chains, which had yet to emerge.
Low liquidity was a chicken and egg problem; the lack of users drove the lack of liquidity. This was before AMMs (Automated Market Maker) were introduced into the space by Uniswap. Augur employed a traditional orderbook model (with decentralized implementation), which resulted in large spreads caused by insufficient liquidity. Another common concern was the volatility associated with markets being denominated in ETH. With their tokens locked up until markets resolved, price swings could render it so that even if users won in a market, they could still end up significantly down in terms of USD. Basically, not everyone who wanted to use the platform necessarily wanted to be long ETH. The obvious solution to this would be use of a stablecoin, but stablecoins were still fairly nascent at the time and Krug cited waiting on MakerDAO’s multi-collateral DAI to come out.
Solving for Unscalabilty
DPM platforms that emerged after Augur improved on some of its issues, but still initially faced roadblocks perpetrated by Ethereum’s long-standing scaling problems. After some failed attempts in earlier years (Veil, Helena), several DPM platforms emerged on the scene in 2020, likely spurred by the DeFi boom. These newer platforms were able to innovate on prediction protocols due to advancements in the Ethereum ecosystem, including AMMs, better stablecoin options, new infrastructure offerings, and emerging scaling solutions.
Omen, launched in July 2020 on Ethereum and powered by the Gnosis conditional tokens framework, improved on the liquidity problem by utilizing an AMM to enable “Uniswap-like” liquidity pools. It also addressed volatility issues by offering markets denominated in stablecoins (in addition to ETH). The use of AMMs in decentralized prediction markets was made possible after Uniswap became the first decentralized platform to successfully utilize them in 2018. However, these innovations didn’t solve transaction speed and gas fees, which only escalated as network congestion increased in the DeFi boom, causing fees to reach unprecedented highs. In February 2021, after months of users asking for a way to use the platform without the exorbitant Ethereum fees, Omen went live on xDai Chain (now Gnosis Chain); an EVM-compatible sidechain with a USD-pegged stablecoin as its native token.
Polymarket followed a similar trajectory. After launching the first stage of its beta on Ethereum in June 2020, the project quickly pivoted to the Polygon network within a few months, citing high friction for users and obscene ETH transaction fees. With the goal of being user-friendly enough to attract non-crypto natives, Polymarket subsidizes users’ transaction fees, a feat more feasible now that it no longer runs on Ethereum. Markets are denominated in USDC and, according to the project’s GitBook, users need only pay gas fees when depositing and withdrawing funds. It is worth noting that, although all trading takes place on sidechains, both Omen and Polymarket, in most cases, still require some interaction with Ethereum when transferring funds to and from the platform. Bridging as well as manual and MetaMask transfers require ETH gas fees and interaction with the network, and the majority of centralized exchanges (where many purchase and cash out tokens) don’t currently support native withdrawals to L2s (although this could change in the future). While the move to L2s has certainly cleaned up the user experience, it hasn’t entirely eliminated the issues.
Augur also took advantage of the advancements in decentralized protocols. In July 2020, they released Augur v2 (now called Augur Pro), proclaiming “when we started Augur, the first 5 years were a lot of research and development and wandering in the wilderness to solve technical problems thought to be impossible. With v2, we’re out of the desert, and now it’s time to execute.” Augur Pro offers markets priced in DAI and an improved UI, but remains on Ethereum and doesn’t utilize an AMM. In The Augur Master Plan (Part 2), the project reported that despite the upgrades, liquidity and usage were still “orders of magnitude” lower than they needed to be. They acknowledged the need to implement an AMM and migrate to an L2. Finally, in May 2021, the team released Augur Turbo, an extension of the protocol that resides on Polygon, utilizes an AMM, and has markets denominated in USDC. This is not an overhaul of Augur Pro, however, which continues to run on Ethereum.
Despite being one of the earliest movers, DPMs have struggled to generate meaningful volume compared to other DeFi verticals. As is evident, this is primarily because user experience is paramount to DPMs and, until recently, they have struggled to deliver on that front. Similar to prediction markets, gaming has been posited as a prime blockchain use case for years. But the popularity of blockchain games didn’t soar until 2021. Why? 2021 was a turning point for them thanks to the progression of L2s and alternative L1s, which provide the throughput needed for a great in-game experience. One of the biggest drivers of Axie Infinity’s (the leading play-to-earn game in 2021) growth was its relocation from Ethereum to its own sidechain. A simplified onboarding process has also contributed to attracting non-crypto native players, another key factor in driving growth.
Too Early
DPMs were simply too early for their time. To provide a great user experience, one that will draw in crypto natives and non-natives alike, there needs to be adequate liquidity, fast and cheap trading, frictionless fiat on/off-ramps, minimal currency-related volatility, and a generally intuitive UI. When the first DPM platform launched in 2018, AMMs had yet to be introduced to decentralized platforms, stablecoins were still in their early days, and developed scaling solutions were a ways off. As other platforms launched in the Summer of 2020, advancements in the space enabled them to improve the liquidity and volatility issues, but it wasn’t until 2021/late 2020 that they were able to begin solving the biggest roadblocks, caused by their underlying infrastructure. Like with gaming, the move to L2s has helped improve user experience a lot. Unlike gaming, which took off last year, the DPM vertical has yet to see explosive growth. DPMs still have some cleaning up to do if they’re going to be competitive with centralized incumbents. The good news is, last year’s breakout adoption of alternative L1s and scaling solutions is bringing that closer to reality.
As increasing emphasis is placed on developing L1 and L2 platforms, growing their ecosystems, and onboarding more users to their networks, the infrastructure needed to optimize the DPM user experience is finally here. In particular, alternative L1s such as Solana, Avalanche, and NEAR, are uniquely poised to solve the issues DPMs have faced, as zero interaction with the congested Ethereum network is required. Solana, for example, offers frictionless fiat on/off ramps (including direct withdrawals to/from centralized exchanges), a throughput of up to 65,000 TPS (~ 4,000x faster than Ethereum), and transaction fees as low as $0.00025. Now you might be thinking, why didn’t the DPM platforms that came after Augur just build on alternative L1s? After all, Solana, Avalanche, and NEAR were all up and running in 2020. While this is true, their ecosystems were not built out at the time. One of the dominant trends in 2021 was the growth of L1 blockchains and their ecosystems. According to The Block’s 2022 Digital Asset Outlook, L1s saw a dramatic increase in quantifiable user activity last year, which was largely driven by the emergence of DeFi ecosystems across the various platforms. While in 2020 nearly all capital locked in DeFi resided on Ethereum, its TVL share has since (as of November 30, 2021) dropped to 63%, as alternative L1s blossomed. The point being, it made sense for DPM platforms that launched in 2020 to go where the users were at the time.
Setting the Stage for Growth
If we apply the Gartner Hype Cycle to decentralized prediction markets, we’re finally in the slope of the enlightenment phase, approaching the plateau of productivity.
As L1 and L2 technology and ecosystems continue to develop, following their outbreak last year, we should see DPM platforms finally gain more traction and user adoption. Improvements to the infrastructure on which they run are crucial to the user experience of DPMs. While this may not have affected the growth of other decentralized platforms as much, it’s clear that DPM users strongly prioritize a frictionless experience over other factors. As the scaling wars wage on, the resulting improvements to blockchains are creating an ideal landscape for DPMs to solve the user experience issues that have plagued them. It won’t happen overnight, but we’re already seeing the shift of new DPM platforms building on alternative L1s, which in addition to being faster and cheaper, can actually support an influx of growth when it comes.
Hedgehog Markets is one such platform. We launched on Solana mainnet in September 2021, and we work a little differently than other DPM platforms. As made evident by the success of Axie Infinity and other play-to-earn games, the P2E model really appeals to users. Hedgehog initially got their start offering No-Loss Competitions, which enable users to participate in prediction markets without the risk of incurring losses. Users stake either a fixed amount of USDC or a Hedgie Hog NFT to enter competitions and are supplied with Game Tokens (which have no monetary value) to use in the markets. Users compete for a top spot on the ROI Leaderboard by making correct predictions, which earns them more Game Tokens, to increase their total game token holdings. At the end of a No-Loss Competition, all users get back their staked USDC or NFT, regardless of performance, and those at the top of the Leaderboard split a USDC-denominated prize pool. This model creates a fun P2E predictions game that both novices and experts can play. The markets on Hedgehog utilize a parimutuel market structure. Our founder, George Yu, made the decision to build on Solana with his primary focus being on creating the best possible user experience. At the time of writing, Hedgehog no-loss competitions have seen upwards of 3,600 unique users stake over $2M in USDC across competitions since its launch. The user experience Hedgehog provides would not be possible without the lightning-fast transaction speeds and near-zero network fees that originate from being built on Solana.
The timing may not have been right in previous years, but it is now! With the infrastructure now in place, and the ecosystem built out enough, Solana is a prime L1 to build a decentralized prediction markets platform that’s able to provide the type of user experience that’s expected. This is why Hedgehog has now launched automated market maker decentralized prediction markets along with Peer 2 Peer fixed odds markets.
In July 2022, we launched our Peer 2 Peer markets, an open sourced program deployed on Solana. We wanted to build a platform where anyone could create, participate, and resolve binary markets on anything. Our Peer 2 Peer markets allow anyone to create a market with whatever odds they choose, select any resolving wallet they would like, and–for one or many wallets–to fund the other side of that market. When a market is resolved, winners will be able to claim their original funds plus winnings. This allows various participants to get paid out without the chosen resolver, or a central person, escrowing and then manually sending out funds to each of the winning wallets.
We are also working closely with on-chain oracles and data feeds to provide a more trustless resolution method. With these feeds users will be able to select from various price, sports and NFT floor price feeds that can be leveraged to resolve a market instead of relying on a person to resolve a market correctly and in a timely manner. We are very excited to be working with some amazing partners and love all of the projects that are being built in the space and on-top of various chains.
Our final push recently was launching our Automated Market Maker (AMM) Prediction Markets. These have been the ultimate goal for us as it allows for community driven price discovery. Some popular markets that this style of market lends itself to are “Will the Ethereum merge to Proof of Stake happen by the end of September 2022?” and “Will the federal reserve announce an interest rate hike of 75bps at the next FOMC meeting?”. AMM style markets can also lend themselves to sports, politics and legislation. With the development of Solana we felt this was a great opportunity to build a DPM that could leverage the speed and minimal transaction fees in order to provide the best user experience.
With the continued developments in DeFi infrastructure, we believe the stage is set for decentralized predictions markets to take off. We are taking advantage of the benefits of the Solana blockchain to provide various market types for users to have the choice as to what suits them best. Bear markets are the best time to build, which is why we launched our Peer 2 Peer and AMM markets and continue to iterate on them by adding more features, design and overall improving the user experience.
While it may not seem obvious just yet, decentralized prediction markets might finally be in the right place at the right time to gain broader adoption. As the DeFi infrastructure continues to develop, DPM’s have the potential to transform political and social processes by involving people in the conversation and rewarding those who can reliably predict.
