Who Actually Wins on Prediction Markets?

Finance-Focused Summary

The video examines who profits from prediction markets, principally Polymarket and Kalshi. These platforms list event contracts—such as political, sports, or weather outcomes—that pay $1 if the specified outcome occurs and $0 otherwise. Users trade against one another through an order book. A contract’s price can be read as an implied probability, though it is not necessarily an objective forecast.

Key Findings and Performance Figures

• Profits are highly concentrated on Polymarket. A study covering November 11, 2022–March 29, 2026, with more than 2.4 million users and $67 billion in trades, found that only 31% of users made money. Among profitable users, the top 1% captured 76.5% of profits, and the top 0.1% received more than half.
• The video’s separate analysis of users overall found that the top 1% captured 88% of profits and the top 0.1% captured 65%. It says 2,480 accounts—about 0.1% of the user base—won over half the money.
• 69% of Polymarket users lost money. Nearly 5%—over 100,000 accounts—lost more than $1,000. 44% stopped trading within a month, with drop-off more pronounced among losing users.
• Skilled winners earned an average of $3,243 over two years in one study; market makers averaged $11,831. The top 0.1% averaged $273,000 in winnings, but traded an average of $16 million in volume.
• A leading election-market winner known as Theo reportedly made $85 million on the 2024 U.S. presidential election after betting more than $70 million. The video attributes his edge to commissioning polls and asking respondents who they thought their neighbors would vote for, rather than how they themselves would vote.
• A weather-market bot built for the video started with $103. It made money on the first two days, then lost money on the next four; after one week, it was down $21.50.
• The video cites $3.7 billion in volume for Polymarket’s main 2024 presidential market. It also says the University of Iowa’s early prediction market beat polls 74% of the time across nearly 20 years of U.S. presidential elections.
• As reported in the video, Polymarket was valued at $21 billion in August 2026, while Kalshi was reportedly raising funds at a $40 billion valuation in June 2026. Kalshi had reportedly been valued at $5 billion the previous October.

Strategies Discussed

The research identified several approaches associated with winning accounts:

• Build an information edge: Gather better or more useful information than the market, as Theo did with commissioned polling.
• React quickly to news: Skilled traders were more likely to trade promptly after announcements, such as Federal Reserve rate decisions.
• Avoid overpriced long shots: Traders may benefit when crowd demand pushes low-probability outcomes above their fair value, while focusing on more likely outcomes.
• Look for arbitrage: In markets with three or more mutually exclusive outcomes, traders may be able to buy every “yes” outcome for less than $1, or every “no” outcome for less than the combined payout of $2 in the example shown. If the pricing conditions hold and execution succeeds, the position has a built-in profit.
• Make markets and capture the spread: Instead of taking an existing quote, a market maker places resting buy and sell orders. In the video’s example, a weather model estimates a 67% chance of temperatures exceeding 80°F, then places orders to buy “yes” at 65¢ and “no” at 31¢. If both fill, the combined cost is 96¢ for a $1 payout, a 4¢ gross spread.
  • This is not risk-free: one side may fill without the other, forcing the trader to take an unfavorable position or pay more to hedge.
  • Effective market making requires many orders and frequent adjustments. The top 0.1% reportedly averaged more than 2,300 trades per day when active, though the median was 78.
• Use a bot only with a sound strategy: Experts caution that automation does not create an edge by itself. A bot needs credible buy/sell signals, reliable data, and execution infrastructure; professional firms may have faster connections and better systems.

Risks, Market Structure, and Context

• Event contracts are described as zero-sum between traders: one user’s profit comes from another user’s loss. After platform fees, the outcome can be less than zero-sum for participants collectively.
• The video argues that major winners tend to have information, trading skill, or substantial capital—and that retail users may be trading against professional market makers or firms. Kalshi has a market-making partnership with Susquehanna International Group.
• Insider trading receives attention, but researchers cited in the video say it does not explain most winning. An analysis by Columbia Law School and the University of Haifa estimated that suspected insiders made $143 million on Polymarket over two years. The video reports that 90% of markets studied were classified as not prone to insider trading, and that the largest identified insider case was not among the top 100 users by profits.
• The Alpha Raccoon account reportedly won over $1 million on a Google year-end search-ranking market priced at under 1%. Authorities allegedly traced the account to a Google information-security engineer and alleged that he accessed the relevant tool before placing bets. As of September 2026, the case was ongoing and the accused had not entered a plea; these are allegations, not a finding of guilt.
• The legal status of prediction markets varies by country and, in the U.S., is contested between federal derivatives regulation and state gambling laws. The video says Polymarket was fined $1.4 million by the CFTC in 2022 and ordered to block U.S. users; Kalshi obtained CFTC approval in 2021 and later prevailed in litigation to list election markets. It also reports that New Jersey asked the U.S. Supreme Court to weigh in in September 2026, with 39 states and the NFL filing briefs supporting the state’s view that sports contracts constitute gambling.
• The platforms also monetize data. The video says Intercontinental Exchange packages Polymarket data for financial firms, while Kalshi has data-distribution arrangements with media outlets and other companies.
• Cautions from interviewees: Don’t assume profits are guaranteed, don’t risk money you cannot afford to lose, don’t trade without a proven edge, and consider stopping after a large gain if you do not believe you have skill. Several experts advise against participation by average users or their own family members.

Assets, Sectors, and Instruments Mentioned

• Platforms: Polymarket and Kalshi
• Instrument: Event contracts—binary “yes/no” contracts, including multi-outcome markets
• Markets and topics: U.S. presidential elections, sports, weather, Google search rankings, chess championships, Elon Musk’s tweets, and Federal Reserve announcements
• Technology and financial infrastructure: Blockchain, smart contracts, central limit order books, and market-making operations
• Firms and organizations: Susquehanna International Group, Intercontinental Exchange, the CFTC, and YouGov
• Not mentioned: Conventional stock tickers, specific ETFs, bonds, or commodity contracts

Disclosures

The video includes an Incogni sponsorship and promotes a discount code. The segment is an experiment and research-based discussion, not a personalized investment recommendation. The experts’ cautions stress the possibility of losing money.

Presenters and Sources

The presenter is not named in the subtitles. Named interviewees and sources include Shane Coplan, Robin Hanson, Vincent Gregoire, Theis (rendered “Tyus” in the subtitles) Jensen, Roberto Gomez Cram, Carissa Veliz, Tarek Mansour, and Luana Lopez Lara.

The video also refers to research by economists at London Business School, Columbia Law School, the University of Haifa, and the University of Iowa, as well as reporting by The Wall Street Journal and The Free Press.