What Are the Chances: The Unlikely Mainstreaming of Prediction Markets

The prediction markets were nearly certain: Francesca Hong would win Wisconsin’s Democratic primary for governor on Aug. 11. Polymarket put the upstart’s chances at 96%; Kalshi was only slightly less sure at 95.5%. 

Then the votes came in. Milwaukee County Executive David Crowley defeated Hong by nearly 4,000 votes. An embarrassed Polymarket immediately deleted an X post from the previous day that had declared, “BREAKING: DSA’s Hong is a near-lock.” Luana Lopes Lara, one of the founders of Kalshi, was quick to note that the upset did not mean the prediction was wrong, because after all, “5% is not 0%.” 

This is true. But Hong’s loss in the face of such confidence was just the latest outcome to cast doubt upon the proposition that betting markets naturally produce better information than governments, media, polls and other institutions. Since this proposition has become a multibillion-dollar industry, its gatekeepers have a clear interest in protecting the myth of what Polymarket founder Shayne Coplan calls a “global truth machine, powered by the people.” But questions about public betting markets’ ability to predict events — never mind debates over the likelihood of massive corruption — are not the only things the industry would prefer the public forget. Far beneath the profiles of industry whiz kids like Coplan and Nate Silver is an origin story that runs through a cast of anarcho-libertarians, proto-Muskean transhumanists, anti-state economists and one of the creepiest carryover figures connecting Iran-Contra and the post-9/11 national security establishment. 

The story begins with an anarcho-capitalist and programmer named Timothy May. After retiring from Intel Corp. in his mid-30s, May spent the late 1980s developing a philosophy based on advances in public-key encryption. Because this technology would enable people to communicate and conduct business anonymously, it opened up the possibility for a new political order he called “crypto anarchy.” As he explained in his 1988 treatise, “Crypto Anarchist Manifesto,” encryption would undermine governments’ ability to tax and regulate economic activity — as well as give rise to anonymous markets for every conceivable area of social and political life. 

“An anonymous computerized market,” he wrote, “will even make possible abhorrent markets for assassinations and extortion. But this will not halt the spread of crypto anarchy.”

In 1992, May joined programmers Eric Hughes and John Gilmore in organizing the first meetings of a group that would become known as the Cypherpunks. They met monthly in Silicon Valley and communicated on an email list that soon attracted more than 1,000 subscribers. Members ranged from civil libertarians concerned with privacy to hardcore anarcho-capitalists like May, who believed encryption heralded the end of taxation, regulation and, eventually, the states itself.

In 1995, a former Intel engineer named Jim Bell brought an extreme proposal to the Cypherpunk list. His 16,000-word essay, titled “Assassination Politics,” proposed establishing an anonymous pool of funds that would pay out to whoever correctly “predicted” the death of government officials. Bell was not interested in using the machinery of markets as a forecasting tool: He saw the assassination payouts as threats weaponized against the state. The more anonymous donors contributed toward a government official’s death — encrypted communications and digital cash would conceal everyone involved — the greater the incentive for someone to make the “prediction” come true. Bell imagined that, with time, no government official would take the risk of seeking and holding office. No officials, no bureaucracy, no government, no problem. 

Bell was convicted in 1997 of harassing government agents and using a fake Social Security number. After serving nearly a year in federal prison, he resumed his feud with the government. In 2000, authorities arrested him again after he was caught conducting surveillance on an IRS agent and his family. Bell was sentenced to 10 years in federal prison on charges that included stalking and intimidating federal agents.

Bell’s crackpot ideas were well-known within the broader libertarian-futurist milieu that anticipated and helped birth today’s billion-dollar prediction market industry. They would find a more palatable expression in Robin Hanson, who joined the Cypherpunk e-list as a CalTech graduate student in the 1990s. In 1991, Hanson published a proposal for an “Idea Futures” in Extropy: The Journal of Transhumanist Thought. It envisioned markets where people could bet on unresolved questions in science and public policy. By aggregating information scattered among many participants, he argued, they would produce better forecasts than traditional institutions. In 1999, his work on prediction markets earned him a post at George Mason University, the nation’s leading intellectual hub of libertarian and free-market economics.

The following year, he was contacted by a program manager at the Defense Advanced Research Projects Agency about applying his work to national security. Hanson agreed, and the result was a project that became known as the Policy Analysis Market, or PAM. The plan was to allow members of the public to bet real money on political, economic and military developments in the Middle East. Shortly after the 9/11 attacks, the project was folded into DARPA’s newly created Information Awareness Office, which was hatched around the same time to develop technologies for identifying and anticipating terrorist threats. Its director, John Poindexter, had served as Ronald Reagan’s national security adviser and was a central figure in the Iran-Contra scandal. Registration for PAM’s first public contracts opened in July of 2003, with trading scheduled to begin that October. 

PAM might have remained relatively obscure were it not for Democratic Sens. Ron Wyden and Byron Dorgan, who held a press conference denouncing it as a Pentagon-funded “terrorism futures market.” They pointed to a sample screen that included bets on the assassination of Yasser Arafat and a North Korean missile attack. The Pentagon killed the program with haste and Poindexter resigned the next day. 

For the next 15 years, prediction markets survived mostly as a niche theoretical interest among libertarian academics.

Then, in 2019, a 21-year-old New York University dropout named Shayne Coplan discovered Hanson’s work. The potential for prediction markets to identify likely outcomes, he later recalled thinking, was “too good of an idea to just exist in white papers.” The following year, he launched Polymarket, raising $4 million from crypto investors including Polychain Capital, Naval Ravikant and Balaji Srinivasan. After the platform correctly called the 2024 presidential election, Coplan believed he and Hanson had been vindicated, proclaiming, “The global truth machine is here, powered by the people.” By then, Polymarket had brought Silver, the election forecaster, on as an adviser. The industry soon exploded: By April 2026, Polymarket and Kalshi were recording nearly $24 billion in combined monthly trading volume. Prediction markets had completed a remarkable journey from the libertarian-futurist fringe to the mainstream.  

Thirty years after Bell imagined anonymous markets around the deaths of public officials, the phenomenon resurfaced in a more respectable form. This year, Kalshi found itself in court over $54 million in bets tied to whether Iran’s supreme leader, Ali Khamenei, would leave office — a question abruptly complicated when his death by a U.S.-Israeli strike paid out $554,000 to an anonymous bettor with the handle “Magamyman.” One can debate the degree to which today’s prediction markets are the direct descendants of Jim Bell’s more brazenly depraved “Assassination Politics,” but the vast array of contemporary public markets related to war and political violence cannot be dismissed as aberrations. They have been central to the concept from the beginning.