Reading Odds as Probabilities

A prediction market price is a probability in disguise. A share that costs 70¢ and pays $1 if correct implies the market thinks the outcome has about a 70% chance. Reading prices this way is the foundation of prediction market trading, but the conversion is not perfect. Spreads, fees, thin liquidity and human biases all distort prices. This lesson shows how to read odds correctly, compare them with your own estimates and decide when a price offers value.

The basic conversion#

implied probability = share price / payout

On a $1 payout, a 70¢ share implies 70%. Yes and No should add up to about 100%: if Yes is 70¢, No should be near 30¢.

Converting from other odds formats#

Adjusting for the spread#

The midpoint between bid and ask is usually the best single estimate of the market's probability, but you pay the ask when buying and receive the bid when selling.

Biases in prediction market prices#

• Long shot bias: people tend to overpay for unlikely outcomes. Studies of betting markets have long found that long shots win less often than their prices imply, while favourites win slightly more often. A 3¢ outcome may really have a 1% to 2% chance.
• Thin markets: a single large trader can push the price away from fair value.
• Partisan or fan bias: people bet on what they want to happen.
• Capital lock up: holding a 95¢ share for months to earn 5¢ ties up money, so near certain outcomes can trade below their true probability.
• Recency and hype: news can cause overreaction. See Recency Bias.

Comparing with your estimate#

The market price is only useful if you compare it with your own estimate.

expected value per share = (your probability × payout) - price

Updating your probability#

Good forecasters update estimates as information arrives, rather than holding fixed views. Bayes' theorem formalises this: start with a prior probability and adjust it based on how strongly new evidence points one way. See Bayes' Theorem and Conditional Probability.

Calibration: checking your skill#

Keep a record of your probability estimates and outcomes. If you are well calibrated, things you call 70% happen about 70% of the time. Group your forecasts into buckets (for example 60% to 70%) and compare with actual hit rates. This is the only honest way to know whether you can beat market odds. See Trading Journal.

Sizing bets by edge#

Even with a real edge, betting too much can ruin you. The Kelly criterion suggests a stake based on your edge and the odds, and many traders use a fraction of it to reduce swings. See Kelly Criterion and Fractional Kelly.

Common mistakes#

• Treating price as truth without considering spread and liquidity.
• Buying long shots because the payout looks big.
• Ignoring time value on long dated markets.
• Never checking calibration.

Frequently asked questions#

How do you convert prediction market prices to probability?#

Divide the share price by the payout. On a $1 payout, a price of 65¢ implies about a 65% chance.

Are prediction market odds accurate?#

They are often reasonably accurate, but spreads, thin liquidity and biases such as overpaying for long shots can distort them.

What is long shot bias?#

The tendency for bettors to overpay for unlikely outcomes, so long shots win less often than their prices imply.

Next, see how the fastest Polymarket markets work in Up or Down Markets Explained.

Sources#

• Wikipedia, Favourite longshot bias