10 Biggest Prediction Market Wins (and What We Can Learn)

In every market, there are trades that become legends. Soros breaking the Bank of England. Paulson shorting subprime. Prediction markets have their own hall of fame — trades where someone saw what the crowd missed, sized up, and walked away with life-changing money. These are the 10 biggest documented wins we could verify. Some come from public leaderboard data. Others from media profiles and court records. A few are composites of multiple traders who made the same play independently.

But before the stories: a warning that we'll expand on at the end. For every one of these winners, thousands of traders made similar bets and lost. You're about to read the highlights reel. The blooper reel is much, much longer.

The largest single prediction market win on record. A French trader identified publicly as "Théo" (later reported as Théo from Paris by the Wall Street Journal) built a massive position on the 2024 US presidential election through Polymarket. He began accumulating contracts months before the election, buying "Trump wins" at prices between $0.33 and $0.40 through multiple wallets.

His total position reportedly exceeded $30 million in contract value. When the contracts resolved at $1.00, his profit exceeded $48 million. He told reporters he'd built a quantitative model using state-level polling data that he believed was more accurate than the market's implied probabilities.

The lesson: Théo didn't have inside information. He had a model he trusted more than the market. And critically, he had the capital and conviction to size his position aggressively when his model diverged from market prices. Most people with the same model would have bet $5,000. He bet $30 million.

The counter-lesson: if his model had been wrong, he'd have lost $30 million. We're only talking about him because it worked.

In January and early February 2020, prediction markets listed contracts on "Will there be 10,000+ confirmed COVID cases in the US by April?" and similar milestones. Most were priced at $0.03-0.10 — the market consensus was that COVID would stay contained.

A small group of traders — epidemiologists, people following Chinese social media, data scientists who'd modeled pandemics — bought aggressively. One documented trader on PredictIt turned roughly $500 into $25,000 by buying pandemic-threshold contracts at $0.05 and watching them go to $1.00 within weeks.

The edge was informational. While mainstream US media was still calling COVID "just the flu" in early February, anyone reading Chinese-language sources or following epidemiological pre-prints could see exponential growth was inevitable. The market was slow to update because most American traders didn't read those sources.

This is the purest example of information edge trading in prediction market history. The information was public. It just wasn't priced because the people trading these contracts weren't the people reading Mandarin-language Weibo posts about Wuhan hospital overflow.

On the night of the Brexit referendum, Betfair had "Remain" priced above $0.85. Polls were close, but the market — and most of the political establishment — assumed Remain would win. Betting volumes on Betfair alone exceeded GBP 100 million, the largest political event in exchange betting history.

Several traders, including at least one documented by the Financial Times, bought "Leave" at $0.15-0.25 in the days before the vote. One trader reportedly turned GBP 15,000 into GBP 110,000. Another, profiled anonymously by Bloomberg, made mid-six figures.

Their edge wasn't polling data — the polls were dead close, around 50/50. Their edge was recognizing that the market was overweighting the "consensus view" and underweighting the polls. When the market says 85% Remain but polls say 50/50, someone is wrong. These traders bet the polls were right and the market's adjustment was the mistake.

The lesson: markets can be anchored by narrative. When every newspaper runs "Remain expected to win" headlines, prediction market traders — many of whom are informed by the same media — anchor to that narrative even when the underlying data says otherwise.

Not a traditional prediction market, but Betfair is an exchange and the mechanics are identical. At the start of the 2015-16 season, Leicester City were 5000-to-1 to win the Premier League. Multiple bettors placed small wagers — GBP 5, GBP 10, GBP 20 — and watched the most improbable sports story in decades unfold.

One bettor, profiled by the BBC, placed GBP 20 at 5000-1 odds and was offered a GBP 72,000 cashout by his bookmaker before the season ended. He took it. (The full payout would have been GBP 100,000.) Others held to the finish.

The lesson here isn't about edge — nobody "predicted" Leicester would win. The lesson is about expected value at extreme odds. At 5000-1, a GBP 20 bet is worth taking even if you think the true probability is 1 in 1,000 (not 1 in 5,000). The mispricing was in the odds themselves. Bookmakers set lines to attract balanced action, not to reflect true probabilities, and 5000-1 on any Premier League team surviving to win is arguably too generous.

Also: the cashout decision. Taking GBP 72,000 versus risking it for GBP 100,000 is rational if you're not wealthy. The marginal utility of GBP 72,000 in hand is much higher than the expected value of the remaining GBP 28,000.

On election morning 2016, Betfair had Clinton at roughly $0.82 and Trump at $0.18. PredictIt had similar pricing. Nate Silver's model gave Trump around 28% — notably higher than the markets — but most traders followed the conventional wisdom.

Multiple traders made significant profits. One PredictIt user, profiled by Vox, reportedly turned $10,000 into $50,000 by buying Trump contracts at $0.15-0.20 in the weeks before the election. On Betfair, aggregate "Trump wins" positions reportedly generated over $1 million in combined profits for contrarian traders.

The interesting wrinkle: many of these weren't professional traders or quants. They were Trump supporters who believed he would win based on rally sizes, enthusiasm gaps, and gut feeling. Their "model" was vibes. It happened to be right. The 538 model, which was more methodical, also suggested the market was underpricing Trump.

This one is hard to draw clean lessons from. Was it information edge? Contrarian courage? Motivated reasoning that happened to land? Probably all three, in different proportions for different traders.

This isn't a single trade but a category. Betfair's tennis exchange allows in-play trading — buying and selling contracts on match outcomes while the match is being played. A small community of professional traders, mostly based in the UK and Australia, have built careers trading tennis matches live.

The documented profits are staggering for individual traders: GBP 50,000-200,000 per year for top performers. They trade hundreds of matches per month, exploiting the fact that Betfair's odds update slower than the actual match momentum shifts. A break of serve changes the match probability significantly, but the exchange price takes 15-30 seconds to fully adjust. That window is where the money is.

This is essentially a latency arbitrage — the same principle as high-frequency trading in equities, but at human speed. The edge isn't in predicting who wins the match. It's in being faster than the market at pricing what just happened.

During the 2022-2023 rate hiking cycle, Kalshi and other platforms listed contracts on specific Fed rate decisions. Several traders made documented profits by reading Fed communications more carefully than the market.

One notable trade: ahead of the June 2023 meeting, the market priced a rate pause at roughly $0.65. But language in the May minutes and subsequent Fed governor speeches strongly signaled a pause, with one governor explicitly saying "skip" in a CNBC interview. A trader buying "pause" at $0.65 collected $1.00 — a 54% return. At scale ($20,000 in contracts), that's $10,800 profit in a few weeks.

The edge was reading primary sources. Most prediction market participants follow financial media summaries. The traders who profited were reading the actual minutes, listening to the full speeches, and parsing the Fedspeak directly.

The SEC's approach to crypto regulation generated dozens of prediction market contracts from 2023-2024. "Will the SEC approve a spot Bitcoin ETF by January 2024?" was one of the most traded contracts on Polymarket, with cumulative volume exceeding $10 million.

Traders who bought early at $0.20-0.30 (when approval seemed unlikely) and held through the Grayscale court victory (which effectively forced the SEC's hand) made 3-5x their money. The Grayscale ruling in August 2023 was the catalyst — anyone who understood the legal implications could see that approval was now 80%+ likely, even as the market was still pricing it at $0.50-0.60.

Estimated profits for early and large buyers: $100K-$500K for traders with five-figure positions. The information edge was legal expertise. Traders who could read the court opinion and understand its regulatory implications moved before the market did.

In the 2019 Australian federal election, every major bookmaker and prediction market had Labor winning. Sportsbet, Australia's largest bookmaker, actually paid out on a Labor victory before election day — they were that confident. Then the Coalition won.

Bettors who'd backed the Coalition collected at odds of roughly 4-1 to 5-1 (implied probability 20-25%). The total payout to contrarian bettors across Australian bookmakers was estimated at AUD 20 million+. Meanwhile, Sportsbet ate the loss on the premature Labor payout AND had to pay Coalition backers.

The lesson: when the market is so confident it's paying out early, that's a signal worth examining. Extreme confidence is often fragile confidence. The polls were within margin of error. The market just refused to believe it.

A less-publicized but instructive trade. In mid-2023, Polymarket listed contracts on whether Russia would renew the Black Sea Grain Initiative. The market was pricing renewal at roughly $0.60. One large trader (visible on the blockchain as a single wallet) bought $50,000+ in "No renewal" contracts at $0.40.

Russia pulled out of the deal in July 2023. The trader's position paid $125,000+ on a $50,000 outlay — a 150% return.

What made this trade noteworthy: the trader appeared to be acting on publicly available Russian government statements. Putin and Lavrov had both signaled dissatisfaction with the deal in terms that anyone following Russian foreign policy could read. But most Polymarket traders were following Western media, which was reporting the deal as "likely to be renewed with modifications." The primary source material said otherwise.

Same pattern as the COVID trades: the information was public, just not in the language or sources most traders were consuming.

Four patterns show up across almost every major prediction market win. First: early conviction. The biggest profits come from buying before the market adjusts. Théo bought months before the election. COVID traders bought in January. By the time a consensus forms, the price has already moved and the upside is gone.

Second: contrarian positioning. Every single win on this list involved betting against the market consensus. That's not a coincidence — by definition, you can't make outsized returns by agreeing with the crowd. The crowd price is the expected value. To beat it, you have to disagree.

Third: information edge from primary sources. COVID traders read Chinese social media. Brexit contrarians trusted polls over narratives. Crypto regulation traders read court filings. Fed traders read the actual minutes. The pattern is consistent: primary sources over secondary summaries.

Fourth: position sizing. Having the right view at small size produces small profits. Théo bet $30 million. The tennis traders make thousands of trades. Edge times size equals profit. Most people with the right view bet too small to matter.

This article profiled 10 winners. It did not profile the estimated 50,000+ traders who made similar contrarian bets on the same events and lost. For every Théo who bet on Trump and won $48 million, there were traders who bet on other candidates at similar prices and lost everything. For every COVID trader who bought at $0.05, there were traders who bought "pandemic contracts" in 2018 and 2019 that expired worthless.

Contrarian bets lose most of the time. That's what makes them contrarian. A contract trading at $0.15 resolves YES only 15% of the time — that's what the price means. The wins are spectacular. The losses are invisible because nobody writes articles about them.

If you're reading these stories and thinking "I should make big contrarian bets on prediction markets," you're drawing exactly the wrong conclusion. The right conclusion is: edges exist, but they're rare, hard to identify in real time, and most people who think they have one don't. Size your bets accordingly. Risk what you can afford to lose entirely. And don't confuse one good outcome with a repeatable strategy.