Prediction Markets: Prices as Probabilities
A prediction market is a venue where contracts pay out a fixed amount if a specific real-world event occurs. The price of the contract — typically between $0 and $1 — is the market’s collective estimate of the probability of that event. Buy at $0.40, the event happens, get paid $1. Buy at $0.40, the event does not happen, lose the $0.40. The price is the probability. This simple structure makes prediction markets unusually clean information machines, because every trader is forced to convert a view into a numeric estimate.
The Three Major Venues
- Kalshi: US-regulated event contracts. CFTC oversight. Covers economic data, elections, sports, weather, geopolitics. Higher friction, higher trust, lower volume on niche topics.
- Polymarket: Crypto-based prediction market. Larger volume on certain topics (elections, geopolitics). Operates offshore. Higher liquidity in headline contracts, regulatory ambiguity.
- Metaculus: Forecasting community, not a money market. Aggregated probability estimates from analysts. Useful as a calibration benchmark, not a trading venue.
Each venue has different strengths. Kalshi for regulated US contracts. Polymarket for global events and high-volume crypto-native trading. Metaculus for clean probability estimates without the friction of price formation.
Why Prices Are Probabilities
A $0.40 contract on “Event X happens” pays $1 if it does. A rational trader buys until the expected value equals zero — at which point the price equals the probability. The math is direct: no discount rate, no growth assumption, no terminal value. The equilibrium price is the market-implied probability. This is the cleanest pricing mechanism in finance.
Base Rates as the Anchor
Most prediction-market mistakes come from ignoring base rates. “Will Country X enter recession in the next 12 months?” has a base rate that depends on macro regime, yield curve shape, and historical frequency. A contract priced far from the base rate without a specific information advantage is usually mispriced against the buyer. Calibration matters more than direction — being right that an event is “likely” without knowing whether the contract is priced at 0.55 or 0.85 is not enough.
Where Prediction Markets Are Useful
Real-time aggregation of distributed information. Faster updates than polls or surveys. A genuine forecast horizon, not a directional bet. For investors: a cross-check on macro narratives that show up in equity moves. When prediction markets and equities tell different stories about the same event, one of them is wrong — and that disagreement is itself a tradable signal.
Where They Are Not Useful
Thinly traded contracts can be moved by a single trader and stop reflecting the underlying probability. Resolution ambiguity can turn a directionally correct view into a losing trade. Some contracts have systematic biases — favoritism toward dramatic outcomes, underweighting of mundane base rates, partisan distortions on political contracts. The signal quality varies enormously across the contract universe.