How Prediction Market Prices Actually Work
Why both sides can be above 50¢ — and why that’s not fraud.
🧮 Quick answer: YES + NO prices don't need to sum to 100¢. The gap is the exchange's spread — the cost of matching two traders who disagree on the probability. Here's why.
Two pricing models — sportsbook vs exchange
The confusion about “both sides over 50¢” comes from mixing up two fundamentally different pricing systems.
Sportsbook (e.g. DraftKings)
- House sets all prices — the book decides what each side is worth
- YES: −110 | NO: −110
- Both sides sum to more than 100% — that extra is the “vig” baked in by the house
- Margin is hidden inside the odds
Exchange (Polymarket, Kalshi)
- Traders set prices via bid/ask orders — no house view
- YES: 65¢ | NO: 62¢ | Spread: 3¢
- Total: 127¢ — that gap is the overround (spread between bid and ask)
- Exchange earns a small fee per trade (0–3%), NOT from the spread itself
* Fee percentages are illustrative. See the platform fee guide for verified rates.
The math: why 65 + 65 = 130, not 100
Step-by-step breakdown of how bid/ask pricing works on a prediction market exchange.
Mid-price ≈ 50¢ (true implied probability). The displayed “65¢” is the best bid — what someone will pay, not what they think the outcome is worth net of spread.
What the displayed price means
“YES at 65¢” means the best available bid — what the highest buyer is willing to pay. It reflects that buyer’s belief that the event has a ~65% chance of happening. It is not a definitive market probability.
The other side of the market
“NO at 65¢” is a different trader who believes the event has only ~35% chance of happening (they want NO at 65¢, implying YES has 35%). These two traders disagree on the true probability — and the exchange matches them.
The spread is the gap between opinions
If the best YES bid is 65¢ and the best NO bid is also 65¢, the mid-price is actually ~50¢ (since YES + NO must eventually pay out to $1 total). The apparent “130¢” comes from reading both bid prices, not the implied probabilities. The spread (3–8¢ on thin markets) is the cost of matching two traders who disagree.
What you actually pay when you buy
To buy YES at market, you pay the ask price — usually a cent or two above the displayed bid. To sell (cash out early), you receive the bid price — usually a cent or two below. That round-trip spread cost is the main reason cash-outs feel lower than expected.
How spread works on each platform
Each platform has a different market structure that determines how wide the spread typically runs.
| Platform | Typical Spread |
|---|---|
| Kalshi | Narrower on liquid markets (politics, Fed rates); wider on niche/low-volume markets |
| Polymarket | Can be wider on sports and thin markets; tighter on high-volume events |
| Robinhood | Same as Kalshi — inherits Kalshi liquidity and order book |
| FanDuel Predicts | Spread not directly visible; margin baked into displayed prices |
⚠️ Spread ranges are qualitative descriptions. Exact spreads vary by market liquidity and are not guaranteed.
The Polymarket tennis example — decoded
A common Reddit post: “I saw YES: 65¢ / NO: 65¢ on a tennis match. That adds up to 130¢. Is this fraud?” Here is exactly what happened.
The scenario
A tennis market shows: YES 65¢ / NO 65¢. At first glance, the totals look like they exceed 100%. So what is going on?
YES buyer at 65¢: “I think Player A wins this match about 65% of the time. I’ll pay 65¢ for a contract that pays $1 if YES.”
NO buyer at 65¢: “I think Player A wins only about 35% of the time. I’ll pay 65¢ for a contract that pays $1 if NO.”
The spread (3¢): These two traders fundamentally disagree. The gap between their bids is the cost of matching them through the exchange. Not fraud. Not a glitch. Just two sides of a two-sided market.
The only time to be concerned
If the mid-price (average of YES bid and NO bid) is wildly different from the true probability — that is a signal worth investigating. PM.us contract comparison tools let you check whether the same event is priced consistently across platforms.
Compare contracts across platforms →Important nuance: On AMM-based exchanges like Polymarket, the displayed price can be the instantaneous execution price, not a resting limit order. On thin markets, the apparent “130¢ total” can spike higher. This is a liquidity issue — an illiquid market is risky, but it is still not fraud.
When does the spread hurt you most?
Not all spread costs are equal. Here are the three situations where the spread has the biggest practical impact on your returns.
Thin / niche markets
Low-volume events — niche sports, local elections, obscure prop markets
Wide spreads on thin markets mean you pay a significant premium to enter AND exit. A 10¢ spread on a $50 position is a 20% immediate drag before any probability movement.
Small position sizes
Trading a few contracts rather than hundreds
Spread cost is roughly fixed per contract. At $10 it hurts more than at $1,000. Always calculate total round-trip spread before entering a small position.
Fast in-and-out trades
Buying and quickly selling before resolution
Every entry and exit pays the spread. A trader who flips 10 times in a week on the same market pays spread 20 times. This is how spread quietly destroys short-term trading edges.
Spread vs fees: These are two separate costs. The spread is the market’s bid/ask gap — it goes to liquidity providers, not the exchange. Platform fees (Kalshi settlement fee, Polymarket trade fee) are charged separately on top of the spread. See the full fee comparison.
Frequently asked questions
Related guides
- Why your cash-out is lower than expected →— path dependence + spread cost ELI5
- Contract comparison across platforms →— are these the same bet?
- How prediction markets resolve →— oracle sources and settlement timing
- Platform fee comparison →— full verified fee breakdown
- PredictionMarkets.US Literacy Hub →— all prediction market guides in one place