Open main menu
Prediction markets may seem new, but betting on the outcome of important events has a long history in politics and otherwise.
Informal prediction markets date back at least a thousand years on myriad events, including betting on the outcome of military battles, on who will become the next king, and on outcomes of the all-important Chinese imperial testing that determined if a man would enter civil service.
More formal prediction markets date back at least five hundred years to the early 1500s in Italy, where markets existed on the next pope and odds would be quoted in correspondence. The first formal "legislation" against prediction markets came in 1591 when Gregory XIV indicated that anyone betting on the outcome of a papal conclave would be excommunicated from the Church.
In Britain, the first documented prediction markets started cropping up at the various London coffeehouses in the 18th Century. Jonathan's Coffeehouse (which eventually became the London Stock Exchange) saw trading in parliamentary scandals and Prime Minister changes starting in the early 1700s. Trading these events became normalized among the elite, and odds would be published in newspapers of the era.
The first documented "whale" sprang up in this environment, Charles James Fox, a member of British Parliament. He was heavily gambling on politics from at least 1771, including betting whether the Tea Act would be repealed. In fact, he likely bet on who would prevail in the American Revolution. He eventually went bankrupt, and his father had to bail him out for tens of millions of dollars (adjusted for inflation). Some may notice parallels to a modern, unnamed American politician.
Betting on prediction markets in America dates back to at least the early 19th Century. James Buchanan, who would later become President, wrote that he lost three tracts of land in 1816 when an election bet didn't go his way. In this era, we also have the first documented American whale: John Van Buren. He was the sitting Attorney General of New York and recorded over 100 bets for a collective $500k (adjusted for inflation) on the 1834 mid-term election. His father, Martin Van Buren (who was himself a documented election bettor in his own right), was Vice President at the time.
Instead of London coffeehouses, the more formal markets in the US centered around New York City pool halls. It was in one pool hall that the first large-scale rules controversy (or to use modern bettor vernacular: rules cuck) happened. The 1876 election was a violent clustercuss. There were more irregularities than a Theranos blood test, and the declaration of a winner was delayed for months. As a result, "Old Smoke" Morrissey, who ran the biggest pool halls in NYC, decided to give everyone back their wagers, with a slight twist: he would keep his commission. He was a renowned boxer and the rival of Bill the Butcher, so I am not sure anyone argued too much about that arrangement.
As in London, odds from New York on the US elections would often be cited in periodicals. At the time, public polling didn’t exist, so betting odds were often the best available indicator of sentiment. In fact, the names of bettors and size of bets would sometimes be tabulated in newspapers. Think of it as the first prediction market leaderboard.
It wasn't until 1936 that Gallup polling would supplant betting odds as the reliable indicator for reporters. Reporting on the odds declined precipitously after that point, and whatever markets existed in New York became taboo in the era during and after World War II, replaced only by informal person to person wagering until decades later.
The advent of bookmaker gambling on elections (the forerunner of peer-to-peer prediction markets) was in the 1960s in London with Ladbrokes. They put up (completely terrible) odds on a conservative leadership race. The 16/1 contender won the race. The tradition of Brits betting regularly and without (much) incident on elections continues into the present with the UK having the largest peer-to-peer betting market in the world, Betfair.
Betfair will put up prediction markets on the highest profile elections and political events, including famously for Brexit. The vast majority of their volume, though, derives from sports. Overall, betting on event outcomes and politics has become normalized in Britain. You could walk into a betting shop in any major city, and place a few quid on just about any major political event you want. If you fancied yourself a great politician, you could also run for office and bet on yourself to win (as more than one candidate has done successfully!).
In the United States, it's been a series of fits and starts. But mostly fits.
The Iowa Electronic Market launched in 1988 as an academic experiment tied to the University of Iowa. The CFTC, the US governmental organization which regulates futures trading, didn't explicitly allow it or disallow the trading, but instead sent them a letter indicating that as long as no person had a position in a market over $500, they would take no action against them. This was the first site to feature 0-100 pricing. If you won, you received $1 per share. If you lost, you received $0. The IEM started small and has purposefully remained small, putting up very few markets with very few users. You would be hard pressed to bet more than a few dollars using their site (let alone $500). Despite still existing, it’s more of a historical footnote than a serious market.
Intrade/Tradesports launched in 2002/2003, funded in part by famous American billionaires Paul Tudor Jones and Stan Druckenmiller. This was a website that offered peer-to-peer binary contracts on the outcome of events. In the same vein as the IEM, the contract was worth $10 if you won, or $0 if you lost. The site was based in Ireland, and had a wink wink agreement with the CFTC starting in 2005 that the US government would not go after their site if they prevented Americans from trading the traditional futures contracts (like the price of gold, oil, and other commodities where trading is heavily monitored and regulated). This became the go-to site for political odds in the 2004, 2008, and 2012 Presidential elections.
The public wasn’t as attuned to data and statistics and numerical analysis during this time period as they are now, so the market for this content was not large as compared to, say, polling. Nonetheless, John Delaney, the indefatigable CEO of Intrade and advocate for prediction markets, was a regular guest on business television like CNBC in the United States, explaining the prices of events on his site, and evangelizing prediction markets more generally. He tragically died climbing Everest in 2011.
Then, a few weeks after the 2012 election, the US government went after Intrade under the auspices that they had violated the 2005 agreement that no Americans could trade oil/gold/other commodity futures. Speaking as a former user who lived in America at the time, I can personally attest that all of these markets would not allow any trades for Americans. But, rather than fight the US government in a years-long and expensive legal battle where even if you won, you would lose, Intrade ejected all Americans and a few months later declared bankruptcy.
Besides launching the first serious prediction markets that allowed US users en masse, Intrade was also known for two famous whales: the McCain whale & the Romney whale. Each amassed gargantuan positions on their respective candidates against Barack Obama. As opposed to the French whale who would appear on Polymarket years later betting very aggressively on a Trump 2024 win, these two bettors lost all of their money, showering winnings on anyone who had Obama.
In 2010, there was a very short-lived market on movie box office futures called the Cantor Exchange (after parent company Cantor Fitzgerald) that got full CFTC approval. I was one of the first people to join. The film industry hated the idea, and lobbied hard against it immediately. The approval lasted less than two months before the US Congress banned it. There are exactly two types of futures banned in the US: onion futures and movie box office (onions because a couple of greedy traders in the 1950s bought almost all of the onion supply in the US, and cornered the market). Fun fact: the long-time head of Cantor Fitzgerald, Howard Lutnick, is now the Treasury Secretary of the United States (this is written in early 2025).
PredictIt – the successor to Intrade in the United States – used the exact same template as the Iowa model: they received a no action letter from the CFTC when they launched in 2014. A US political consultancy company called Aristotle teamed up with Victoria University out of New Zealand to create the site. They capped each position at $850 (which was the IEM limit, adjusted for inflation). PredictIt became the most-cited source of odds for the 2016 and 2020 elections. The CFTC retracted the no action letter for PredictIt in 2022, after PredictIt began to put up markets that the CFTC felt veered too closely into gambling. This included the number of tweets that politicians would publish each week. One such market, "How many tweets will Andrew Yang post this week?" led to called-in threats against Andrew Yang to tweet more that eventually involved law enforcement. Ultimately, in its first incarnation, PredictIt flew too close to the sun in trying to extend the bounds of the CFTC remit, and paid the price. There are now rumors of PredictIt trying to retool, rebrand, and relaunch as a regulated site, if the CFTC gives their blessing.
2020 saw the launch of the current-era prediction markets: both Polymarket and Kalshi. Both were still very minuscule in 2020.
2020 also featured other cryptocurrency websites with crude prediction markets. Multiple entrants split up the minimal Presidential volume: Augur, Catnip, and FTX to name a few. On these sites, you bought crypto tokens that sat in your crypto wallet and were worth $1 if your candidate won and $0 if they lost. All of these markets were supplanted and replaced by Polymarket, and barely exist except as jokes or memories. Notably, FTX got into some slight trouble with the law for unrelated reasons. Side note: one of my friends is (theoretically) owed so much money from a Trump 2024 bet that SBF put on the Alameda books that it became a line item in the bankruptcy documents.
As of 2025, prediction markets are dominated by two players, Kalshi and Polymarket. They technically exist with two user bases who can’t interact with one another, which often leads to slightly different event prices across the two markets.
Kalshi is a website/app with hundreds of markets. After getting a kickstart in Y Combinator, Kalshi received full CFTC approval in 2020 for an events-based market with the exclusion of elections. After formally being denied election contracts in 2022 by a CFTC commissioner vote, Kalshi sued the CFTC. In 2024, after a Supreme Court case ended Chevron deference and severely weakened agencies like the CFTC, a judge ruled in Kalshi's favor that they could list election contracts. In 2025, Kalshi pushed the legal envelope even further by allowing betting on sporting events from an event contract standpoint. They are currently being sued by multiple US states. Kalshi does not allow non-Americans, and all bets are denominated in dollars.
Meanwhile, Polymarket is a cryptocurrency-based website with hundreds of markets. All transactions are on-chain, all bets go through a smart contract, and all of the smart contracts go through a third party verifier called UMA before paying out. Polymarket paid a fine to the CFTC in January of 2022 and agreed to ban Americans from using their site. While they do use cryptocurrency, the bets are denominated in a stablecoin called USDC, run by Circle. Fluctuations in cryptocurrency prices do not impact user's bets. The website is located on an ethereum L2 called Polygon. To translate that to English using a crude analogy, Polymarket is like an app, Polygon is like the operating system, and Ethereum is like the manufacturer of the phone.
Both companies were started in New York by highly-ambitious founders that are almost the same age. Each company is very well-capitalized and has a long list of well-known investors and consultants. The rivalry could be likened to Uber versus Lyft or Visa versus Mastercard.
Each company’s volume, and prediction market volume in general, exploded in the 2024 election as political chaos in the United States (especially the unprecedented meltdown and late drop-out of a party’s presidential candidate) led the media to focus more than ever on betting odds to make sense of the world.
One story that particularly captured the public's attention was my discovery and conversation with the "French whale", a shadowy gentleman who became the single largest prediction market bettor in history over the course of a few weeks. He bet tens of millions on Trump to win the 2024 election, and moved the odds sharply into Trump’s favor. In the end, he was right.
The future of these markets is likely to be shaped by each company trying to become the dominant player in the space. The likely outcome, however, is that both companies will be behemoths in 2028, with smaller companies trying to take pieces of the pie.
Regulatory issues remain a headwind in the US, especially in light of Kalshi pushing the legal envelope on sports, and in light of Polymarket’s founder being raided by the Justice Department at the tail end of the Biden presidency.
As volume has exploded on prediction markets in recent years, the demand for new and innovative markets is higher than ever. This is both a boon and a challenge: rules issues are now the biggest hindrance to the industry. And this will inevitably delay the next huge growth spike for markets as new users become frustrated by the need for byzantine expertise in navigating the issue.
Overall, prediction markets, despite being killed off again and again in history, continue to sprout up given any daylight. When two people disagree about an important issue, there’s no better way to settle it than putting your wallet behind your words. And synthesizing all of these words and wallets into a large market unlocks wisdom.
The future is unknowable and unpredictable.
Do not think of prediction markets as an all-important solution to figure out what is going to happen, but instead as a better flashlight to look ahead to what the future might be. Markets are certainly better than punditry. And they’re even better than polling, because markets incorporate polling.
And as the population is becoming more polarized in western democracies, with people getting news from increasingly biased sources, prediction markets can and will cut through the partisan bullshit and pay off the truth.
