A Prediction-Market Price Is an Estimate With Frictions

A prediction-market quote can be read as an estimate of an event's likelihood, but the number is also a transaction price. It reflects the contract's rules, available buyers and sellers, and the costs of taking a position. Calling a 62-cent price a 62% probability can be useful shorthand, provided those qualifications remain attached.

The first question is what the contract pays. The CFTC describes event contracts that often have two outcomes and a fixed payment, usually $1, when the specified outcome occurs. Other structures exist. An explanation based on a simple binary contract should not be applied unchanged to a product with partial payouts or multiple outcomes.

Start with a payoff, then calculate the result

Assume a hypothetical Yes contract costs $0.62 and pays $1 if a stated event occurs, otherwise zero. Before fees, the successful outcome produces a $0.38 profit. The unsuccessful outcome loses the $0.62 purchase price. The advertised $1 payment includes the money originally put at risk; it is not an additional dollar of profit.

Now add an illustrative $0.02 purchase fee. Total initial cost becomes $0.64. A successful outcome leaves $0.36 profit and an unsuccessful one loses $0.64. If a reader independently assessed the event's probability at 67%, the hypothetical expected profit would be $0.03: 67% of $1, less the $0.64 cost. That personal probability estimate could be wrong.

Expected value is an average across imagined repetitions under an assumed probability. A single contract still resolves according to its own outcome. A positive calculation is neither a guarantee of profit nor evidence that the analyst's probability estimate is accurate. Fees can also vary by platform, transaction type and contract.

The wording defines the event

A contract about a published economic statistic must identify the statistic, reference period and source used for settlement. A first estimate can differ from a later revision. A contract that settles on the initial release can therefore have a different result from one using the latest revised figure, even if their headlines sound similar.

Imagine a hypothetical question asking whether an index exceeds 100 in December. A value exactly equal to 100 would not satisfy a strict greater-than condition. A different contract saying at least 100 could resolve the other way. Date cutoffs, rounding rules and the specified official series can matter as much as the broad economic view.

This is why a reader should separate the underlying real-world development from the settlement test. A business may clearly have improved while failing to meet a narrowly worded contract condition. The contract pays on that defined condition, rather than on the most persuasive narrative about the event.

A screen number may not be an executable price

Suppose a market's most recent trade was at $0.70, but current buyers bid $0.60 and sellers ask $0.64. A new buyer would normally need to consider the available ask and quantity, rather than assume a purchase at the last-trade price. Buying at $0.64 and immediately selling at $0.60 loses $0.04 per contract before fees.

A larger order may use up the best available quote and receive worse prices for the remaining quantity. The number displayed in a headline can therefore differ from the effective price for a specific trade. An apparent jump in probability in a thin market deserves particular care if it comes from only a small transaction.

The crowd is not a representative survey

Participants are self-selected and have unequal budgets, information and reasons for trading. Some may be hedging an exposure elsewhere. A company facing a possible loss from an event might accept an unattractive standalone price because the contract provides protection for that business. Its purchase need not express the same forecast as an unhedged speculator's purchase.

Trading volume also counts transactions rather than unique independent opinions. The same position can change hands repeatedly. A busy market may incorporate useful information, but turnover alone does not prove that a quoted probability comes from a broad or representative sample of people.

Use the price as evidence with boundaries

A useful comparison records the contract, time, bid, ask, available size and settlement rule. It then compares that market evidence with the relevant data or a clearly specified forecast. Claims about current regulatory status require separate checking; this explanation does not resolve the jurisdictional questions discussed in TTN's legal-background article.

The number becomes more informative when its limitations are visible. A prediction-market price shows what participants can currently exchange a defined payoff for. It can contribute to a forecast without becoming an objective measurement of the future or a substitute for understanding the event itself.