How do prediction-market odds work?

Price ↔ probability

Prediction-market shares trade between 0¢ and 100¢, and that price maps directly onto a percent chance: 63¢ means the market is pricing the event at 63%. Buy a YES share at 63¢ and the event happens, you collect the full $1.00 — a profit of 37¢ per share. If it doesn’t happen, the share expires worthless and you lose the 63¢ you paid. There’s no separate “odds” notation to convert — the cent price already is the probability.

YES and NO are mirror images

Every market has two sides, and they have to add up to (roughly) $1.00 between them. If YES trades at 63¢, NO trades at roughly 37¢ — buying NO at 37¢ is the exact same bet as saying “this event won’t happen,” just expressed from the other side. You can take either side of any market; there’s no structural advantage to being the YES buyer versus the NO buyer.

Spread and liquidity

The order book has a best bid (highest price someone will pay) and a best ask (lowest price someone will sell for). The gap between them is the spread. A tight spread on a deep book means the displayed price is a solid read on the crowd’s probability estimate. A wide spread means the market is thin — few participants, light volume — so that “probability” is fuzzier, and a single sizeable order can move it a lot more than it would on a liquid market.

Why prices move

Prices shift as new information arrives — a poll, a news story, an earnings report, a game score — and as traders reposition around it. They also tend to drift toward the extremes (0¢ or 100¢) simply as a market’s resolution date approaches and outcomes become clearer, since less genuine uncertainty is left to price in. A market sitting at 50¢ with weeks to go and one sitting at 50¢ the day before resolution represent very different amounts of real uncertainty.

Are these probabilities any good?

On balance, yes — large, liquid prediction markets have historically tracked real-world outcome frequencies fairly closely, which is the whole reason people treat the price as a probability in the first place. They’re not perfect, though: like other betting markets, prices can show a favorite-longshot bias, where long-shot outcomes (very low cent prices) trade a bit rich relative to how often they actually happen, and favorites trade a bit cheap. We publish live calibration data so you can see how well prices have actually tracked outcomes, rather than taking that on faith.

Key terms

Implied probability

The share price expressed as a percent chance — a 63¢ share implies a 63% probability. It’s the market’s current collective estimate, not a guarantee.

Spread

The gap between the best bid and best ask. Narrow spreads signal an actively traded, liquid market; wide spreads signal a thin one.

Bid / Ask

The bid is the highest price a buyer currently offers; the ask is the lowest price a seller currently accepts. You buy at the ask and sell at the bid.

Favorite-longshot bias

A well-documented pattern in betting and prediction markets where long-shot outcomes are priced slightly too high and heavy favorites slightly too low, relative to how often each actually occurs.

Expected value

What a share is worth, on average, if your own probability estimate for the event differs from the market price. It only means anything relative to your own honest estimate — the market price is not a target to beat, it’s the baseline you’re comparing your view against.