How Prediction Market Prices Map to Probabilities
Binary contract prices translate directly into implied probabilities. Here's how to read them — and where the simple translation breaks down.
Every binary prediction contract pays $1 if the event happens and $0 if it doesn't. If the market is pricing the contract at $0.62, traders collectively believe the event has roughly a 62% chance of happening. This direct mapping — price equals probability — is what makes prediction markets so legible. It's also why journalists cite them as odds even when the underlying trading is thin.
The basic math
Expected value = (probability of yes) × $1 + (probability of no) × $0. If the market price equals expected value, the price IS the probability. A risk-neutral trader who thinks the true probability is 70% will buy any Yes contract priced below $0.70 and sell any priced above. The bid/ask spread is where market-makers collect a small edge for providing continuous liquidity.
For mutually exclusive outcomes (Yes and No), the two prices should sum to exactly $1 in a frictionless market. In practice they sum to slightly less than $1 because of spread — the difference is the market-maker's edge.
Where the translation gets noisy
Fees, spread, and time value all push price away from the true probability. On Kalshi, a fee on profitable closes means a fair Yes price is slightly above the true probability for Yes-buyers, because you'll pay a fee on any winning trade. On Polymarket, a wide bid/ask spread on illiquid markets means there's no single "price" to read — the mid is a rough guess.
Long-dated markets also discount for opportunity cost — capital tied up for a year should earn a return, so contracts on far-future events trade below their true probability by roughly the risk-free rate. If T-bills yield 5% and a market resolves in a year, a $0.50 contract implies about a 52.5% true probability once you account for the time value of money.
Worked example: reading an aggregated headline
Suppose a news article reports "Polymarket gives Candidate X a 63% chance to win." Peek at the actual order book. If Yes trades at $0.62/$0.64 with $2M of depth on each side, the 63% is a reasonable midpoint and there's real conviction behind it. If Yes trades $0.55/$0.71 with $50k of depth, the 63% headline is essentially made up — the market is thin enough that a single $10k order could move the price 5 cents.
The spread tells you how confident the market is in its own number. A market quoting 60/64 is far less certain than one quoting 61.8/62.2. Journalists rarely include the spread; sophisticated readers always check.
Reading aggregated odds
When media outlets cite "a 62% chance according to Polymarket," they're reading the midpoint of the bid/ask on the flagship market. That's a fine first approximation for headline purposes but it hides the depth and the spread. For serious use, look at the actual book, or better, look at time-weighted average price over the last day to smooth out noise.
Common pitfalls
Don't over-interpret a $0.90 price as certainty — a 90% market still fails 10% of the time. Don't compare prices across platforms without checking the resolution criteria; a $0.60 Kalshi contract and a $0.62 Polymarket contract may be pricing subtly different events. Don't confuse implied probability with your own probability — that's the source of your edge, or your loss.
When to trust prices as probabilities
Trust them most on deep, liquid, actively-traded markets with unambiguous resolution — flagship elections, Fed decisions, major sports. Trust them less on thin markets, novelty markets, and markets with subjective resolution. As a rule of thumb: if a $10,000 order would move the price by more than a cent, treat the current price as a rough estimate rather than a precise probability.