The map of participants, what each earns, and what that means for retail traders.
Forbes reported $24B/month in prediction market volume as of mid-2026. That’s volume — not profit. Here’s the full map of who actually captures value in that volume: the exchange collecting fees, market makers earning spreads, informed traders exploiting information advantages, and retail participants navigating all three.
Understanding this map doesn’t just explain where the money goes — it tells you which type of participant you are, what edges exist for retail, and where the structural headwinds come from.
Every trade involves at least two sides. But behind every market are four distinct participant roles — each with a different earnings mechanism and a different relationship to risk.
Every trade generates a fee for the exchange — regardless of who wins or loses. Fee structures vary significantly by platform and market type.
| Platform | Fee model | Effective rate |
|---|---|---|
| Kalshi | Per-contract taker fee (formula-based, entry only) | Up to 1.75¢ per contract (near-zero for heavy favorites; politics/policy: zero fees) |
| Polymarket | Probability-based taker fee (category-dependent) | Sports 0.75% peak; Crypto 1.80% peak; Politics/Finance/Tech 1.00%; most fee-free at extremes |
| FanDuel Predicts | 2% of potential payout at checkout (same rate on early exit) | 2% of potential payout at checkout |
| ForecastEx (Interactive Brokers) | Exchange fee built into contract price | $0 commission; $0.01/contract exchange fee built into price |
Fee rates change. Confirm current rates at each platform’s official fee schedule before trading. See our fees comparison for a full breakdown across platforms.
A casino takes the other side of every bet — the house edge means the casino profits from your losses directly. Prediction market exchanges don’t do this. They earn fees on volume, not on outcomes. Your wins and losses flow between other participants, not to the exchange. The exchange earns the same whether you win or lose.
A market maker posts simultaneous bids and asks — for example, 41¢ bid and 43¢ ask on a YES contract. When a buyer transacts at 43¢ and later a seller transacts at 41¢, the market maker has earned the 2¢ spread. They take on inventory risk in between: if the market moves sharply, they can be stuck holding the wrong side.
In liquid, high-volume markets (major elections, World Cup), competition among market makers compresses spreads to 1¢ or less. In thin markets, spreads can reach 5–10¢ — a significant hidden cost. On a 40¢ contract, a 6¢ round-trip spread is a 15% cost before the outcome is even decided.
What this means in practice
Checking the bid-ask spread before entering a position is as important as checking the contract price. A wide spread on a small position can erase the expected return from a correct call. Liquid markets are cheaper to enter and exit — but they’re also the markets where professional flow is deepest.
“Informed trader” is not a synonym for insider. Most information-edge trading is legal — and understanding the distinction matters both for evaluating your own edge and for understanding who you’re trading against.
Primary-source speed
Reading regulatory filings, official results, or primary data sources faster than the market prices them in.
Quantitative models
Building statistical models that outperform naive crowd pricing — particularly on base rates, historical analogies, and calibrated probability.
Domain expertise
Deep knowledge of a specific sport, political system, economic indicator, or geography where crowd pricing reflects general assumptions, not expert analysis.
Trading on material non-public information (MNPI) — stolen data, access to unpublished results before they are released, or information obtained through breach of duty — is prohibited by exchange rulebooks and subject to CFTC enforcement action. Exchanges monitor for unusual position concentration ahead of resolution events.
See the PM insider trading enforcement trackerWin conditions for retail
Expert flow is sparse; crowd pricing is less informed. A local political race or regional sports event may price incorrectly because professional traders focus elsewhere.
Before market makers arrive in volume, pricing can be wide and inaccurate. Getting in early — at a price that later corrects — is a real edge. Thin markets cut both ways.
Automated price-updating is not instantaneous. If you read a primary source faster than the market prices it in, you can trade ahead of the adjustment.
Common loss patterns
Major elections, high-profile sports finals — these attract the deepest professional flow. Retail trading in these markets means facing the best-informed traders with the fastest execution.
In thin markets, a 5–8¢ spread on a 40¢ contract is a 12–20% round-trip cost before the outcome is even decided. Many retail losses are really spread costs, not bad calls.
If a YES contract reaches 95¢ before the event resolves, holding to 100¢ earns 5¢ per contract — while the position was exposed to reversal risk the entire time. Selling at 94¢ is often rational.
See Can retail traders win on prediction markets? for a deeper look at the structural conditions where retail has genuine edge, and why prediction market traders lose money for a detailed breakdown of the loss mechanics.
In poker, one player wins the pot after the house takes a rake from every pot. The house is structurally always profitable. In prediction markets, no entity takes the other side of every trade. Outcomes flow between participants — the exchange takes fees on volume, but it doesn’t collect your losses directly.
Like stock exchanges, prediction market exchanges earn fees on volume. Multiple participants interact: market makers provide liquidity, informed traders supply price information, and retail participants add volume. The key difference from equities: PM contracts are binary, short-duration, and expire at a fixed settlement.
The key structural difference from both
Prediction market contracts expire. Unlike holding a stock indefinitely, every PM position resolves at a binary value (1 or 0 per contract). There’s no waiting for a recovery — when the event concludes, positions settle. This creates a different risk profile than equities and a different time horizon than poker hands.
Gambling vs. investing: the full comparisonYour returns on prediction markets depend on four factors:
Which market you choose
Thin markets have wider spreads but less professional competition. Liquid markets have tighter spreads but deeper expert flow. Neither is automatically better — it depends on what edge you bring.
What edge you bring
Research depth, domain expertise, or faster access to primary information are the sources of durable edge. Opinion is not edge. Being right for the wrong reasons doesn't compound.
What you pay in fees and spreads
Exchange fees and bid-ask spreads are real costs that compound across trades. A correct call can still be a losing trade if the entry and exit costs exceed the return.
Whether your information advantage is legal
Legal information edge (research, expertise, primary sources) is permitted and widespread. Material non-public information is banned. The line is about the information's source and the duty it carries, not its accuracy.
Does the platform take the other side of my trade?
No. Prediction market exchanges don't act as counterparty. They match buyers and sellers, then collect a fee. If you buy YES, another user sold YES to you. The platform earns the same fee whether you win or lose.
What is a market maker and why does their spread matter to me?
A market maker posts a buy price (bid) and a sell price (ask) simultaneously — for example, 41¢ bid / 43¢ ask. When you buy at 43¢ and later sell at 41¢, the 2¢ difference went to the market maker. In large, liquid markets (major elections, World Cup finals), competition among market makers compresses spreads to 1¢ or less. In thin, low-activity markets, spreads can reach 5–10¢ — a significant hidden cost on small positions.
When do retail traders actually win?
Retail participants have a genuine edge in three specific conditions: (1) niche or local events where professional flow is thin and crowd pricing is naive; (2) entering illiquid markets before market-maker flow consolidates; (3) reacting to breaking news faster than the market's automated pricing responds. Most retail losses happen in the opposite scenario: liquid, widely-followed markets where professionals have better data, faster models, and can move size without moving the price.
Is this like poker with a house rake?
Closer to stock-market trading than to poker. In poker, one player wins the pot after a rake is taken. In prediction markets, multiple participants interact — the exchange takes fees, market makers earn spreads, and informed traders edge retail. There is no single 'house' taking the other side of every contract. The exchange profits from volume, not from your loss.
Can Retail Traders Win on Prediction Markets?
The 4 conditions where retail has genuine edge — and where you're structurally outgunned.
Why Prediction Market Traders Lose Money
The mechanics of the most common loss patterns, from spread costs to market selection.
How Concentrated Is Prediction Market Profit?
Who holds the bulk of winning positions — and what it implies for the average trader.
Fees Comparison
Side-by-side fee comparison across all major prediction market platforms.
Insider Trading on Prediction Markets
The difference between legal information edge and insider trading — and how enforcement works.
Gambling vs. Investing: Is This a Bet or a Trade?
The structural distinctions and what they mean for how you should approach a position.
