The platforms look like simple probability apps. The economics increasingly look like an electronic market where casual flow meets specialists, market makers and automated traders.

The $1 Contract Is Becoming a Trading Instrument

Prediction markets used to be discussed mainly as forecasting tools: put money behind a belief, aggregate everyone’s information, and the market price becomes a probability. That description is still useful. It is no longer sufficient. Robinhood reported 13.6 billion event contracts traded in the second quarter of 2026, more than ten times the prior-year level; July alone reached 6.1 billion contracts and August 4.7 billion.  [1] [2] [3]

At that scale, the useful comparison is no longer only with polling. Prediction platforms are becoming electronic trading venues with order books, maker and taker economics, market-making programs, APIs, institutional OTC desks, liquidity incentives and cross-venue arbitrage. A retail user sees a binary question. A specialist sees a price, a spread, a probability distribution, a settlement rule and a queue.

Figure 1. Robinhood illustrates how quickly event contracts have moved into mainstream retail distribution.

Prediction Markets Are Usually Not “The House vs. the Customer”

The first misconception is that every losing customer dollar becomes platform profit. On an exchange-style prediction market, traders are normally matched against other traders. The platform monetizes the transaction layer rather than taking a directional view on the outcome. The CFTC describes regulated prediction exchanges as outcome-indifferent venues, and Kalshi explicitly says participants trade against another member, not the exchange.  [4] [5]

That makes the basic economics closer to an exchange than to a sportsbook. If Yes trades at $0.60, the holder is paying $0.60 for a claim that settles at $1 if the event occurs and $0 otherwise. Ignoring costs, the opposite side funds the complementary exposure. The platform can then earn from commissions, transaction fees or related services, while liquidity providers compete to capture spread and incentive revenue.

Who Makes the Money? The Data Point to a Barbell

The clearest large-sample evidence comes from Polymarket’s onchain history. Galaxy Research analyzed 2.9 million accounts classified as retail and found 69.2% below break-even. The cohort was down $338.9 million in aggregate. A separate group of roughly 125,000 accounts classified as automated finished up $246.8 million. Galaxy notes that many automated accounts are reward-churning bots, while a smaller subset performs genuine market making and arbitrage.  [6]

Figure 2. The two cohorts are not a complete zero-sum partition, but the direction of the split is striking.

This does not mean the typical user is being financially destroyed. Pew Research Center examined a shorter six-week period and found the average Polymarket trader spent a little more than $600 and had a net loss below $2; 58% gained or lost less than $100. Seven percent earned more than $1,000, while 9% lost more than $1,000.  [7]

The two studies measure different windows and populations, but together they suggest a barbell. A huge number of casual participants appear to use prediction markets in small size, while meaningful profits and losses concentrate in a much narrower tail. That is exactly what should be expected when recreational flow meets participants who specialize, automate and trade repeatedly.

The “Sharps” Are Not Merely Better at Guessing

A specialist’s advantage can come from several layers at once. The obvious one is information: a trader who follows one policy process, sports league, technology niche or economic release more closely may update probabilities faster than a casual participant. But modern prediction-market edge increasingly looks like trading infrastructure rather than trivia knowledge.

The Financial Times recently profiled a private group of highly specialized Polymarket traders whose members use data analysis, rapid intelligence gathering and collaboration to trade geopolitical and political markets; the group has reportedly generated tens of millions of dollars collectively. That is anecdotal, but it is consistent with the broader shift toward professionalization.  [8]

Institutional infrastructure is arriving too. In June, Galaxy launched an OTC prediction-market desk and disclosed a $10 million Kalshi trade with crypto hedge fund Arca. Galaxy can warehouse principal risk and hedge across prediction venues and traditional markets—capabilities far beyond the workflow of a casual app user.  [9]

Market Makers Can Earn the Spread — Until Information Jumps Through It

Prediction-market makers perform the same basic job as liquidity providers elsewhere: post two-sided prices and earn a spread when other participants cross it. Kalshi now operates a designated market-maker program whose approved participants can receive reduced fees and adjusted position limits in exchange for ongoing quoting and volume obligations. Polymarket charges no maker fee and uses taker fees on selected categories to fund daily maker rebates; it also supports separate liquidity-reward programs.  [10] [11] [12]

But event contracts contain unusually sharp adverse-selection risk. A market maker quoting 48/52 around a 50% event can earn pennies repeatedly—until a verified result, injury, court ruling, data release or official announcement causes fair value to jump almost instantly toward 0 or 1. The trader who hits the stale quote may know more than the liquidity provider. That means the spread is not free income; it is compensation for inventory risk, jump risk and the possibility of trading against better information.

In thin markets, the problem becomes more severe. Galaxy has noted that some low-liquidity prediction markets can move by 10 percentage points or more on only a few thousand dollars of flow. A displayed 63% “probability” in such a market can therefore be partly a statement about the latest order rather than a stable consensus estimate.  [13]

Fees Quietly Raise the Probability You Need to Be Right

Because the payout is capped at $1, small transaction costs can consume a surprisingly large share of a modest statistical edge. Polymarket’s current fee schedule charges takers on selected categories using a probability-weighted formula, while makers are free and may receive rebates. Finance and politics markets use a 4% fee parameter; sports uses 5%; crypto uses 7%; geopolitical and world-event markets are currently fee-free.  [11]

At a 50-cent price, 100 contracts in a Polymarket finance market generate a $1 taker fee. The position itself costs $50, so the entry fee alone equals 2% of the capital committed. If the trader exits before resolution, another trading fee may apply. A trader who believes true probability is 52% therefore does not necessarily possess a usable 2-point edge after spread, fees and estimation error.

Robinhood’s retail event contracts make the same point in a different way. Its platform commission is probability-weighted and capped at one cent per contract, while the underlying exchange can charge an additional fee of up to one cent per contract. For 100 contracts around a 50-cent price, the Robinhood commission reaches the $1 cap; the exchange charge could add up to another $1.  [14]

Longshots Reveal the Entertainment Layer

The strongest evidence that some participants are buying excitement rather than probability is the favorite-longshot bias. A September 2026 academic study analyzed 588 million Polymarket trades across 2.48 million accounts. In its transaction-flow specification, purchases below 10 cents lost 19.3 cents per dollar committed, while purchases at or above 90 cents earned 0.83 cents per dollar. The magnitude changes under alternative grouping methods, so the result should not be treated as a universal law—but the aggregate longshot penalty is economically meaningful.  [15]

Figure 3. Very cheap contracts can resemble lottery tickets: low upfront cost, huge headline payoff and poor aggregate pricing for buyers.

The psychology is intuitive. A five-cent contract can return $1—a 20× gross payout—if the unlikely event occurs. That framing makes the upside salient and the small purchase price psychologically disposable. The same behavior appears in lotteries and sports betting. Prediction markets add a financial interface, a live chart and a probability label, but those features do not automatically remove entertainment demand.

Galaxy’s account-level work finds a related topic effect: sports specialists were the least profitable specialization in its dataset, while technology and science specialists were the most profitable. Galaxy cautions that the latter group may include subject-matter experts and, in some cases, traders with unusually strong information advantages. The safest interpretation is not that one category is inherently easy, but that expertise matters more when casual flow is abundant.  [6]

The Platform Has a Different Business Model From the Sharp

The venue does not need to forecast better than its customers to make money. It needs activity. Kalshi says its revenue model is transaction fees rather than taking positions on outcomes. Polymarket now charges taker fees in selected categories and recycles part of those fees into maker rebates, while some categories remain fee-free. Robinhood charges its own event-contract commission on top of possible exchange fees.  [16] [11] [14]

This creates a familiar brokerage incentive: engagement and turnover can matter more than customer direction. A participant who repeatedly buys and sells may be economically valuable to the platform even if the user’s long-run P&L is flat. Liquidity programs can also subsidize activity during the growth phase. Kalshi currently operates both liquidity incentives and a volume cashback program for eligible users, while Polymarket permits sponsors to fund market-specific liquidity rewards.  [17] [18] [19]

A Prediction Price Is Not the Same Thing as an Objective Probability

A 67-cent price is often displayed as “67%.” That is a useful translation, but retail users should understand what is hidden inside it. The price is the marginal clearing level produced by current participants, current liquidity, current fees, position limits, settlement rules and current information. In a deep market with diverse participants, that may be an excellent probability estimate. In a thin market dominated by a few correlated traders, the label can create more confidence than the order book deserves.

This matters especially when a market becomes entertainment-driven. If a fan base systematically overpays for a beloved team, a political constituency trades expressively rather than probabilistically, or social media sends a crowd into a low-price longshot, the market is still discovering a price—but part of that price may be demand for participation rather than information.

Regulation Can Improve Market Integrity Without Creating an Investor Edge

U.S. regulated event contracts sit inside the CFTC derivatives framework. The CFTC emphasizes exchange surveillance, anti-manipulation rules, customer-fund protections at intermediaries and transparent contract terms. It has also been updating its prediction-market framework as the category expands; approximately 1,600 event contracts were certified for listing in 2025, compared with an average of roughly five a year from 2006 through 2020.  [4] [20]

Those protections matter, but they do not turn a retail participant into a profitable forecaster. Regulation can reduce fraud, improve rule clarity and police manipulation. It cannot eliminate adverse selection or stop a better-informed counterparty from taking the other side of an ordinary trade. The CFTC’s September advisory on “mention markets,” for example, specifically highlights heightened manipulation risk when settlement can depend on an individual’s discrete conduct.  [21]

The Retail Checklist: Are You Trading Probability or Buying Participation?

What Changes as the Market Professionalizes?

Prediction markets are likely to become harder for naive directional traders if institutional liquidity and automation continue to expand. Better market makers should tighten spreads. Cross-venue arbitrage should eliminate obvious pricing discrepancies more quickly. AI agents can monitor thousands of contracts and update fair values continuously. Those are all improvements for market quality—but they also mean easy mispricings disappear faster.

The professionalization can therefore help retail users and hurt them at the same time. Execution can improve while alpha becomes scarcer. A narrower spread reduces the cost of being wrong, but a better-priced market also reduces the reward for being casually right. In mature financial markets, retail investors generally do not beat professional market makers by reacting to public headlines a few minutes later. Prediction markets are moving in the same direction.

Conclusion: The Product Is Probability; the Business Is Flow

Prediction markets can be excellent forecasting mechanisms, useful hedging instruments and compelling entertainment products at the same time. The mistake is assuming those identities produce the same economics for every participant.

The exchange monetizes activity. The market maker monetizes spread and liquidity incentives while carrying jump risk. The sharp monetizes information, modeling, execution and rule interpretation. The hedger may willingly accept negative standalone P&L because the contract offsets a larger external exposure. And the casual trader may rationally pay a small amount for engagement—but should not confuse that utility with positive expected return.

For retail users, the most important shift is conceptual: stop asking only “What do I think will happen?” and start asking “What probability am I being offered, what probability can I justify, and who is likely to be on the other side?” Once prediction markets become real trading venues, that is where the money is made.

Methodology & Data Notes

• The Galaxy profitability figures are based on its October 1, 2026 analysis of Polymarket onchain activity. Galaxy defines and filters “retail” and “automated” cohorts; the two P&L totals should not be treated as a complete zero-sum accounting identity for every wallet and every platform flow.
• The Pew figures refer to a six-week observation window and therefore measure a different horizon from the Galaxy full-history analysis. The difference is useful: short-window typical outcomes can be small even when long-run aggregate losses are material.
• The favorite-longshot results are from a September 2026 academic preprint using 588 million Polymarket trades. The quoted -19.3 cents and +0.83 cents per dollar are from observed transaction flows; the authors report that magnitudes vary with how related contracts are grouped.
• Robinhood contract counts are counts of event contracts bought or sold, not dollar notional. Each contract settles at $1 or $0. July/August 2025 values are reported rounded in Robinhood’s comparative disclosures.
• Fee examples are illustrative and use published schedules current as of the research date. Platform fees, exchange fees, rebates and incentive terms can change and can vary by category and account tier.

Sources

1. Robinhood Q2 2026 Results

2. Robinhood July 2026 Operating Data

3. Robinhood August 2026 Operating Data

4. CFTC — Understanding Prediction Markets and Event Contracts

5. Kalshi — Who Are You Trading With?

6. Galaxy Research — The Behavior of Polymarket Traders

7. Pew Research Center — What We Know About the Typical Polymarket User

8. Financial Times — Polymarket’s Alpha Traders

9. Galaxy — Institutional OTC Prediction Markets Trading

10. Kalshi — Market Maker Program

11. Polymarket — Trading Fees

12. Polymarket — Liquidity Rewards

13. Galaxy Research — The Shape of Prediction Markets to Come

14. Robinhood — Event Contracts Overview and Fees

15. Cardozo & Rivero-Wildemauwe — The Favorite-Longshot Bias in Prediction Markets

16. Kalshi — How Does Kalshi Make Money?

17. Kalshi — Liquidity Incentive Programs

18. Kalshi — Volume Incentive Program

19. Polymarket — Sponsor Market Rewards

20. CFTC — 2026 Prediction Markets ANPRM

21. CFTC — September 2026 Mention-Market Advisory

Johan Shamshad

Financial Markets Analyst and Journalist at Dave Finances |  More PostsBio ⮌

Johan Shamshad is a financial markets writer at Dave Finances covering cryptocurrencies, trading platforms, brokers, fintech, financial regulation, and developments across global markets. He previously worked at Gulf News, adding newsroom experience to his coverage of fast-moving financial and digital-asset markets.

His work focuses on identifying market-moving events, company developments, regulatory changes, product launches, and shifts in trading and financial infrastructure.

Johan contributes news and analysis designed to help readers understand not only what happened, but why a development matters and how it may affect the wider financial landscape.

• Johan Shamshad

  The Trading Boom Changed Retail FX — But Who Actually Bears the Risk?
• Johan Shamshad

  Tokenized Treasuries Explained: Why Crypto Investors Are Suddenly Buying Government Debt
• Johan Shamshad

  What Happens When Stocks Start Trading Like DeFi Tokens?
• Johan Shamshad

  Stablecoins Are Starting to Look Like Savings Accounts — But They Aren’t Bank Deposits