How prediction markets work

A plain-English explainer · Updated August 2026

Prediction markets let people buy and sell shares in the outcome of a future event. The price of a share becomes a real-time, crowd-sourced estimate of how likely that outcome is. Here's the machinery behind that number — and how to read it well.

New to the topic? Start with what prediction markets are for the big picture; this page digs into the mechanics.

Prices are probabilities

Most markets settle each share at $1 if the outcome happens and $0 if it doesn't. So if "Will Team X win the title?" trades at $0.40, the market is implying roughly a 40% chance. As new information arrives — an injury, a trailer drop, a poll — traders buy and sell, and the price moves to reflect the new consensus. That is why a market price is often a surprisingly good, if imperfect, probability estimate. A handy shortcut: multiply the price by 100 to read it as a percentage, and remember the "Yes" and "No" prices of a clean market should add up to about $1.

Two ways markets are built

Under the hood, platforms match buyers and sellers in one of two ways, and it helps to know which you're looking at:

• Order books (used by exchanges like Kalshi) work like a stock exchange: buyers post bids, sellers post asks, and trades happen when they meet. The "price" is the last trade or the midpoint of the best bid and ask.
• Automated / central-limit order-book crypto exchanges (like Polymarket) settle in stablecoins on a blockchain, but the reading is identical — a share price between $0 and $1 that maps to a probability.

Either way, the takeaway is the same: nobody hands down the odds; the odds are whatever people are currently willing to pay.

Buying, selling, and closing early

You don't have to wait for an event to resolve. Because shares trade continuously, a holder can usually sell at the current price at any time — taking a profit if the price has moved their way, or cutting a loss if it hasn't. This is a key difference from a traditional bet slip, and it's why a market price reacts to news minute-by-minute rather than only at the finish line.

How a market resolves

Every market names a resolution source and a deadline up front — the official result, a data release, a named authority. When the event concludes, the market "resolves": the winning shares pay $1 and the losing shares expire at $0. Well-designed markets spell out the exact criteria to avoid ambiguity ("resolves YES if the official CPI release shows…"). Occasionally a messy real-world outcome leads to a dispute over how a market should settle, which is one reason clear, high-volume markets are more reliable than vague ones.

Why volume and liquidity matter

A market with heavy trading volume has more people, more money and more scrutiny behind its price, so its number tends to be more meaningful than a thinly-traded one. Liquidity — how easily you can trade without moving the price — goes hand in hand with volume. In a thin market, a single large order can knock the price around and make it look like sentiment shifted when really one trader just showed up. Prediction Headlines ranks each category by volume for exactly this reason: the busiest markets are usually the most informative and the most talked-about.

Why prices stay honest: arbitrage

If "Yes" and "No" ever add up to much less than $1, or the same event is priced differently in two places, traders can lock in a near-riskless profit by buying the cheap side — and that buying pushes the prices back into line. This constant hunt for mispricing is part of what keeps a liquid market's number well-calibrated. It also breaks down in thin markets, where the profit isn't worth the effort, so odd prices can linger.

Single-outcome vs. multi-outcome

Some questions are simple Yes/No ("Will it happen by date X?"). Others have many possible winners (a league champion, an award). For multi-outcome markets we show the current favorite and its implied odds; the full field is always a click away on the source market. In a well-formed multi-outcome market the probabilities of all the candidates should add up to roughly 100%.

Common mistakes when reading the odds

• Treating a probability as a prediction. A 70% favorite still loses about three times in ten. The market is telling you the odds, not the result.
• Trusting a thin market. Low volume means the price can be one enthusiast's opinion, not the crowd's.
• The favorite–long-shot bias. Very unlikely outcomes are often priced a little too high and heavy favorites a little too low, because people overpay for lottery-like bets.
• Ignoring the clock. "By end of year" and "this week" are different questions; always check the deadline.

What prediction markets are good — and bad — at

They shine at aggregating dispersed information and reacting fast to news. They struggle when a market is illiquid, dominated by a few traders, driven by hype, or asking about something genuinely unknowable (which is where our "Oddball Bets" come from — entertaining, low-probability questions about aliens, the end of the world, and the like). Used well, a market price is one of the best single-number forecasts you can get for free; used carelessly, it's just a vibe with a dollar sign.

Are these bets?

The underlying platforms are trading venues that may or may not be available where you live. Prediction Headlines itself is purely informational: we summarize public data and link out. We take no bets and hold no funds.