How Prediction Market Odds Work: Reading Prices as Probabilities

Prediction market odds are prices. A YES contract on a question like "Will the Fed cut rates at its next meeting?" trades somewhere between $0.01 and $0.99, and that price is the market's implied probability: YES at 34¢ implies roughly a 34% chance the event happens. Every contract settles at exactly $1 if the event occurs and $0 if it does not, so reading the odds is as simple as reading the price. Everything else — order books, spreads, fees, cross-venue divergence — is market structure layered on top of that one mechanic.

What does a prediction market price actually mean?

Binary event contracts are quoted in cents, and the cents map directly to percentage points. YES at 34¢ reads as a 34% implied probability. YES at 80¢ reads as 80%. Because YES and NO are two sides of the same event, their prices are complementary: if YES trades at 34¢, NO trades near 66¢, and the pair sums to roughly $1 before spread and fees.

This is the core difference from sportsbook odds formats. There is no moneyline, no fractional notation, no decimal conversion. The price is the probability, updated continuously as traders buy and sell. If you are new to the underlying concept, start with what a prediction market is — this post assumes that baseline and focuses on how to read and interrogate the numbers.

How do YES and NO contracts settle?

Every binary contract resolves to one of two values: $1 if the event occurs as defined by the market's rules, $0 if it does not. That fixed settlement is what makes the arithmetic clean.

Work the numbers on a single contract:

• Buy YES at 34¢ and the event happens: the contract settles at $1. Profit is about 66¢ per contract before fees — roughly a 194% return on the 34¢ at risk.
• Buy YES at 34¢ and the event does not happen: the contract settles at $0. The full 34¢ is lost, and never more than that. Maximum loss is always your entry price.
• Buy NO at 66¢ and the event does not happen: settlement at $1 yields about 34¢ profit per contract before fees.

Scale is linear. One hundred YES contracts at 34¢ cost $34 and pay $100 at settlement if the event occurs — $66 of profit before fees, or a $34 loss if it does not.

You are not locked in until resolution, either. Contracts trade continuously, so a trader who bought YES at 34¢ can sell at 55¢ after the odds move and bank about 21¢ per contract before fees, without ever taking resolution risk to the finish line. One caveat that catches newcomers: markets settle on their written resolution criteria, not on the headline you think you traded. Read the rules before the trade, and check how a contract is defined when you browse live events.

How do venues set prices — order books or AMMs?

Two mechanisms dominate. Most major venues today, including Kalshi and Polymarket, run central limit order books: traders post bids and asks at prices they choose, and trades print where buyers and sellers cross. The "price" you see is typically the last trade or the midpoint between the best bid and best ask. Order books give tight pricing when market makers are active, but a thin book means wide spreads and real slippage on size.

The alternative is an automated market maker (AMM), the design many crypto-native venues launched with — Polymarket itself started on an AMM before moving to an order book. An AMM replaces human market makers with a liquidity pool and a pricing formula: every buy pushes the price up along a curve, every sell pushes it down. AMMs guarantee you can always trade, but the price impact of a large order is baked into the math rather than negotiated in a book. The distinction still matters across the long tail of crypto prediction markets, where AMM-style designs remain common.

For reading odds, the practical takeaway is the same either way: check depth, not just price. A 34¢ quote backed by deep resting liquidity is a much stronger statement than a 34¢ quote that a $200 order would move five points.

Is the market price the same as the true probability?

No — and this distinction is where most sloppy prediction-market commentary goes wrong. The price is the market's best tradable estimate, which is not identical to the true underlying probability, for several structural reasons:

• Spread and fees. If YES is bid 33¢ and offered at 35¢, there is no single "market probability" — there is a band, and your effective entry sits on the expensive side of it.
• Capital lockup. Money in a contract is committed until settlement or exit. A YES priced at 97¢ on an event that resolves in eight months may look "cheap" relative to a near-certain outcome, but the return on locked capital can make buyers scarce, holding the price below a pure forecast.
• Risk preference and hedging. Some participants buy contracts as insurance against outcomes they fear, not as forecasts, and they will pay above fair value for that hedge.
• Longshot bias. Betting-market research has long documented a tendency for low-probability outcomes to trade rich relative to how often they actually happen.

Deep, liquid markets tend to be well calibrated on average; thin or emotionally charged ones drift. The evidence on when market prices are trustworthy — and when they are not — is worth its own treatment, which we cover in are prediction markets accurate.

The expected-value framing follows directly from the arithmetic. If YES trades at 34¢ and your own estimate of the probability is 45%, your expected value is 0.45 × $1 − $0.34 = about 11¢ per contract before fees. The trade is not "will this happen" — it is "is the price wrong."

How do spreads and fees compare to sportsbook vig?

A sportsbook hides its margin inside the odds. The classic two-sided line at −110/−110 implies about 52.4% on each side, so the two implied probabilities sum to roughly 104.8%. That extra ~4.8 points is the vig — a built-in cost you pay no matter which side you take, with no way to trade out of the position.

Exchanges unbundle the same cost into visible parts:

• The spread. If YES is 33¢ bid, 35¢ offered, crossing the book costs you about a penny versus the 34¢ midpoint. The exchange analog of overround appears when the YES ask and the NO ask sum to more than $1.
• Explicit fees. Venues charge differently — trading fees, settlement fees, or withdrawal costs vary by platform and change over time, so check each venue's published schedule rather than assuming.

The structural advantages of the exchange model are that the cost is visible, that patient traders can post resting orders and earn the spread instead of paying it, and that positions can be exited mid-event. The fee and structure differences between the two venues W.E.T. tracks most closely are laid out in Kalshi vs Polymarket.

How do prices move when news hits?

Odds move when information moves, and most tradable information arrives on a schedule: CPI prints, Fed decisions, court rulings, debates, earnings, protocol upgrades. These are catalysts, and prediction market prices behave around them the way options behave around known events — positioning builds beforehand, spreads widen into the release as market makers pull back, and the print itself gets repriced in seconds to minutes.

The move is itself the message. A market that jumps from 34¢ to 60¢ on a headline is telling you, in one number, how much that headline changed the probability of the outcome — which is often more informative than the headline. That is the core W.E.T. editorial model: pair the story with the odds move, the volume, and the next catalyst on the calendar. Our market news feed is built around that framing, and the prediction markets guide covers the catalyst lens in more depth.

Fast markets deserve respect. Around a major release, resting orders go stale, spreads gap, and market orders can fill far from the last print. If you are trading the catalyst itself, limit orders are the difference between a price and a hope.

Why do the same odds differ across venues?

Kalshi and Polymarket regularly price near-identical events a few points apart. The causes are structural: different user bases, different fee treatment, different collateral rails (regulated US dollar accounts versus crypto collateral), different geographic access, and — most importantly — different resolution rules and sources.

The textbook arbitrage works like this: if YES trades at 40¢ on one venue and NO trades at 55¢ on another for the same event, buying both costs 95¢ and returns $1 at settlement no matter what happens — about 5¢ of gross profit per pair. In practice, fees, capital sitting locked on two platforms until resolution, and the risk that the two contracts do not resolve identically eat much of that edge. A gap is only true arbitrage if the resolution criteria genuinely mirror each other; a wording difference can turn "risk-free" into a two-sided loss.

Even untradable divergence is signal. When two independent pools of traders disagree by several points, one of them is wrong, and figuring out which is a research prompt. Tracking that spread across venues in one place is exactly what the event dashboard is for.

Where should you go next?

Once price-as-probability clicks, the rest of prediction markets is application: find the catalyst, judge the liquidity, compare the venues, and decide whether the price is wrong. Watch live odds and cross-venue divergence on the event dashboard, and compare notes with traders working the same markets in the W.E.T. community. Everything here is informational, not financial advice — prediction market trading carries real risk of loss, and the only odds that matter are the ones on your own screen.