There was no better time to launch a new crypto-based predictions market than in 2020. The uncertainty of the U.S. presidential election and the continually shifting narratives around the coronavirus pandemic called out for new ways for society to come to consensus.
That’s the founding story Shayne Coplan, the 22-year-old creator of Polymarket, has gotten into the habit of telling. His platform is just the latest in a long line of cryptocurrency-based predictions markets. But it's also a rare example of a crypto-project finding early "product-market fit."
“We're growing and we've got traction,” he said. Less than a year after founding, and after at least one major code overhaul, Polymarket has grown to do more than $100 million in total volume to date. “That was organic. No s**tcoin, no subsidies, no bots. People just wanted to use this product.”
Crypto is full of projects based on sound ideas, stable code and deep pockets that struggle to get off the ground – often because they attempt to reinvent what already works. The projects that seem to take hold, like Bitcoin and Ethereum, are usually the ones that execute a simple idea, simply.
Polymarket allows users to open a betting market on any subject and launch it as a tamper-proof smart contract on an Ethereum layer 2. Backed by Naval Ravikant, Balaji Srinivasan, Meltem Demirors as well as a number of other crypto-focused VCs and builders, it has an air of legitimacy for a project attempting to do what many others have struggled to achieve.
Some of the earliest, like Augur, and largest, like Gnosis, initial coin offerings were to build crypto-based prediction markets. Augur famously launched after three years of development with a clunky product that only recently upgraded to something more useable. Gnosis has shifted gears to a community-run project after failing to live up to expectations. Other betting markets, like Veil, simply shut down.
"Frankly, no one had pulled it off," Coplan said. "These legacy incumbents who had raised tons of money have never shipped anything usable." That may be an overstatement, but not by much. Polymarket's strong growth and expanding community has been attributed to its sleek, low-friction design – but it also may be from being in the right place at the right time.
Today, there are questions on whether President Biden will still be in office by month’s end, whether Floyd Mayweather will KO Logan Paul, as well as bets on future coronavirus caseloads. There's a lot to be uncertain about, a boon for a product that pitches itself as an antidote to misinformation.
New markets are started all the time, with many of them seeing hundreds of bets worth hundreds of thousands of dollars. Some people seem to be looking for consensus opinion about controversial topics, others are gambling on news events – like you might see in any pub.
“Both people who are looking for an edge and people who are looking to make better decisions are our North Star for now,” Coplan said. A firm believer in markets theory, Coplan thinks prediction platforms are a genuine way to come to a better understanding of reality.
The idea is rooted in economic theory, which Coplan knows inside and out. In a winding interview with CoinDesk, the curly-haired New York University dropout made continual references to legendary Austrian school economist Friedrich Hayek.
The thinking goes, when there are economic incentives at play, people are more likely to give an accurate read of uncertain possibilities. They read more and better sources, think deeper and attempt to place their money on whatever realistic outcome is more likely to occur.
“As the coronavirus happened, and there was just so much uncertainty and so many differing opinions, [I thought] ‘If only there were free markets on these topics, so people could put their money where their mouth is,’” Coplan said. This attitude is foundational to the crypto ethos, which distrusts established sources of authority and places a high-degree of responsibility on individual actors.
Robin Hanson, associate professor of economics at George Mason University, has tracked the phenomenon for the better part of his career, establishing himself as one of the foremost economists researching prediction markets. “People have noticed that speculative markets have had this information-aggregation benefit,” he told CoinDesk. “That's an observation that goes back centuries.”
See also: How Crypto Transforms Prediction Markets
Hanson’s primary concern aligns with Coplan’s: determining ways that price action could reveal obscured truths. “The world has just broken in terms of how people come together to share information and draw conclusions.”
Where Coplan and Hanson separate is in their willingness to adopt blockchain for this purpose. Coplan thinks prediction markets have been “the Holy Grail” for the cryptocurrency industry since inception. Both crypto and prediction markets can trace their lineage to neoclassical economics, which championed free markets and individual rights.
By disintermediating betting markets from centralized players – which could inherently bias a prediction either by determining what questions people are allowed to ask, or otherwise interfering with the outcome – Coplan thinks Polymarket, and its “copycats,” are finally realizing this ideal.
“That's like the one true use case for blockchains that really makes sense. It's creating this Mecca of real time, unbiased, perpetually accurate information that anyone in the world can go and reference to make better decisions,” Coplan said.
Gwern, an independent AI researcher and long-time prediction markets (PM) user (though not Polymarket), agreed. “Decentralized PMs seem critical in the long run for any kind of widespread adoption,” he said over a direct message. “They're too vulnerable to trusted third party security vulnerabilities: step on too many toes, and speak too much truth to power.”
Earlier experiments with crypto-based betting markets, like Augur and Gnosis, were torpedoed by Ethereum’s long-standing scaling problems, he said. Further, both initially launched with a native token, adding friction to the user experience. Polymarket learned from its predecessors' mistakes.
Last autumn, Coplan led a team of 10 coders to transpose his platform onto Matic, an Ethereum layer 2, to lower the gas fees involved in placing bets. He also designed the system for users without an ether wallet. All bets are placed in the dollar-pegged stablecoin USDC, which can be purchased on Polymarket with debit or credit cards, through an integration with Transak.
“We grew too quickly and we've encountered a number of problems based on the underlying infrastructure,” Coplan explained.
The ease of use, as well as the high-profile of many of the platform’s backers has drawn a dedicated user base to Polymarket.
A group of mostly anonymous users who call themselves Polymarket Whales track trades on the platform to aggregate on an independent website and broadcast over Twitter. Coplan calls them “professional truth seekers,” though it’s likely they’re amateur researchers, potentially with a stake in many of the open markets. (Polymarket Whales did not respond to multiple requests for comment.)
Looking closely at the preserved data on the blockchain sometimes reveals the odd ways prediction markets make sense of the world. There was a bet on whether Kanye West and Kim Kardashian would file for divorse before Jan. 1, 2021 – a prediction that ultimately missed the mark by about a month, but points to the predictive ability of crowds.
Coplan thought that divorce market was “irrelevant” – just Page Six gossip – but pointed to examples of where Polymarket helped uncover useful information. During the Suez Canal crisis, which saw a lodged tanker block access to one of trade finance’s arteries, whales were pulling information about wind speeds, Egyptian river flows and other “seemingly arbitrary details” that helped people build a better understanding of the situation.
Like any market-based system, elements of greed and unethical games can often seep in. LessWrong, a rationalist blog, detailed what looked like a Tesla insider placing a bet on whether the auto manufacturer would buy bitcoin just days before the pivotal event occurred. It was apparently his or her first trade, and a large gamble at that. They pocketed $120,000, data shows.
This type of activity has Hanson, the economist, wondering about the actual societal benefit of platforms like Polymarket. To him, browsing the site, it seemed like it was missing the point of what a well-formed predictions market could provide. Namely, predictions are supposed to help inform actual behavior – otherwise you’re running numbers for the mob.
“Those are the kinds of information that are valuable,” he said. “But merely generating information on a topic where nobody is taking substantial actions is not actually very valuable. There's no point.”
See also: Martin Glazier – Enterprise Blockchain Doesn’t Work Because It’s About the Real World
Hanson said Polymarket should add conditional variations on the questions people can raise. Rather than framing whether an event may happen with a binary yes or no, there would be more value for users if they could ask about likely outcomes based on changing circumstances. (Coplan says conditional markets could be in the offing.)
Coplan acknowledges that markets are only fit to answer some types of questions. “Truth is oftentimes probabilistic rather than definite,” he said. “The ability to draft a market on something is a good litmus test for how legitimate it is.”
There could be other reasons for participating in a predictions market, Hanson added. It could just be plain fun. It could also be a decent way to make a buck. Gwern, who only ever bet low amounts, said it was mostly for intellectual stimulation.
The frivolity of today’s decentralized predictions markets may be their saving grace, Hanson said. Historically, decentralized prediction markets sold themselves as a way to bypass existing laws, he said. Nonprofit, centralized markets PredictIt and the Iowa Electronic Market, for instance, operate in the U.S. under the legal protection of a no-action letter from the Commodity Futures Trading Commission (CFTC).
“They do have a lot of regulations available to them to enforce,” he said, but have mostly turned a blind eye because it’s a nascent industry that hasn’t done much harm yet.
But because Polymarket doesn't book revenue – Coplan gave a non-committal response when asked if he was thinking about monetizing it – and because it never launched a token, it might become an accepted part of the internet’s open architecture.
It sounds like the type of probabilistic questions Polymarket was set up to ask.