What are prediction markets? How event contracts turn prices into forecasts

Prediction markets is trading venues where people buy and sell contracts that pay out based on a future outcome, and the resulting price is treated as a crowd forecast. That “forecast” is only as good as the market’s incentives and constraints, including who is allowed to trade, how liquid the contract is, and what regulators permit.

Key Takeaways

• Prediction markets trade an event contract, often a binary contract that pays $1 if a defined outcome happens and $0 if it does not.
• The contract price is commonly read as an implied probability, but that interpretation weakens fast when liquidity is thin or informed traders are restricted.
• U.S. access depends on the regulatory wrapper: Kalshi operates as a CFTC-regulated Designated Contract Market, while PredictIt relied on a narrow no-action framework and Polymarket faced a 2022 CFTC enforcement action.
• “Most accurate platform” rankings are fragile because accuracy scoring and market conditions change the results more than most comparisons admit.

Prediction markets and event contracts

A prediction market is a market for claims on an outcome. The thing being traded is an event contract whose payoff is tied to a clearly specified real-world event, like an election result or an economic release. Many venues list these as a binary contract: a “Yes” share that pays a fixed amount if the event happens and a “No” share that pays if it does not. That contract design is why the price is easy to interpret and why prediction markets show up in headlines as “odds.”

If a contract pays $1 on “Yes” and trades at $0.63, the screen is telling you the market is willing to clear risk at 63 cents for a dollar of conditional payoff. Traders translate that into an implied probability of roughly 63% because the payout is fixed and the price is bounded between $0 and $1. That translation is useful, but it is not magic. It is a derivatives price with a settlement rule, fees, spreads, and positioning behind it.

This is where most explainers stop, and it is where traders start. The price is a forecast only to the extent the venue can attract informed flow and let that flow express itself. If the market is thin, a small order can move the price far away from what a larger, deeper pool of participants would clear. If the venue bans the most informed participants, the price can become a popularity contest instead of an information aggregator.

The “prediction markets crypto” angle matters here because crypto rails make it easy to list and trade these contracts globally, but they do not remove the need for a credible settlement process and a regulatory perimeter. A market can be on-chain and still be a low-quality signal if it cannot reliably settle, cannot attract informed money, or cannot legally serve the traders who would correct mispricings.

How prices become crowd forecasts

Three mechanics turn a pile of opinions into a single number: incentives, price updating, and resolution. Incentives are straightforward. If a trader believes the true chance of “Yes” is higher than the current price implies, buying “Yes” has positive expected value. If enough traders act on that edge, their orders push the price toward their estimate.

Many prediction markets use a market scoring rule to update prices as trades arrive. Robin Hanson, a George Mason University economist, is cited as helping develop the market scoring rule used by many prediction markets and argues that insiders moving prices toward the truth is central to the model’s accuracy. The key point is not the formula. The key point is that the mechanism is designed to pay for information by letting informed traders profit from correcting the price.

Resolution is the hard part, and it is where “how do prediction markets work” becomes operational. Every contract needs a rule for what counts as “Yes,” what data source is authoritative, and what happens if the outcome is ambiguous. That is the plumbing behind how prediction markets resolve. Some venues rely on centralized adjudication under a rulebook. Others lean on crypto-native dispute systems and oracles. In crypto markets, a common pattern is an optimistic oracle design where a proposed outcome stands unless challenged within a window, with disputes arbitrated under predefined rules. The uma optimistic oracle is one well-known example of that optimistic pattern.

Once resolution is credible, the price can function as a forecast. Without credible resolution, the price is just a number attached to a promise. That is why the quality of the signal depends on constraints. A market that cannot attract informed participation, cannot absorb noise with liquidity, or cannot settle cleanly will produce a price that looks like an implied probability but behaves like a thinly traded option.

Where prediction markets are used

The obvious use case is forecasting public events. WealthManagement frames prediction markets as trading event contracts, often binary yes or no, whose prices are used as forecasts of real-world outcomes. That framing matches how these markets are consumed. People look at the price as a live, continuously updated estimate that incorporates new information faster than a weekly poll or a static analyst note.

A second use case is decision support, where the market is not just predicting the world but helping choose an action. Manifold describes a futarchy setup using conditional prediction markets: define a set of possible actions and a success metric that will be known later, create a market for each action predicting the metric conditional on taking that action, pick the action with the highest predicted success at decision time, revert trades on the non-chosen actions, and later resolve the chosen market when the metric is known. Manifold’s 2025-09-30 post describes a public experiment applying this structure to select AI safety research projects for the MARS program, using defined binary success metrics and a selection rule based on predicted outcomes.

This matters because it changes what “accuracy” means. In a pure forecasting market, the goal is calibration against reality. In a decision market, the goal is to pick the best action under uncertainty, which can reward traders for identifying second-order effects and hidden constraints, not just guessing the headline outcome.

It also changes the incentives around participation. If a decision will be made based on the market, the market becomes a target for persuasion, lobbying, and strategic trading. That does not automatically break it. It does mean the reader should treat the market as a mechanism embedded in a social and regulatory environment, not as a neutral oracle.

Accuracy, incentives, and manipulation risks

Accuracy talk gets sloppy because people mix three different questions: whether prices are well-calibrated, whether the market is liquid enough to be meaningful, and whether the participant set contains the people who actually know things. Manifold’s community discussion on 2026-05-06 shows how fragile cross-platform comparisons are. It lists a snapshot ranking (Metaculus 92%, Polymarket 89%, Manifold 77%, Kalshi 74%, Augur 59%, Betfair 47%, PredictIt 34%, Iowa Electronic Markets 34%) while explicitly warning that comparisons depend on methodology and market conditions like liquidity and trader count.

The comments make the same point in trader language. One participant argued that “surely something like https://calibration.city/ is a more objective measure,” pointing at calibration-style scoring rather than vibes. Another said, “Manifold is definitely not the most accurate, but it is the most convenient,” which is a blunt reminder that UX and accessibility can drive participation even when it degrades the signal.

The insider-trading controversy is not a side issue. Fortune reports platforms including kalshi and polymarket adding restrictions that bar politicians from trading on their own campaigns, athletes from trading in their own leagues, and employees from trading on contracts tied to their employers. The same Fortune piece quotes Hanson arguing insiders should trade because they move prices toward the truth and that the purpose of the market is to inform decisions. That is the core tradeoff. Restricting insiders can reduce ethical and operational risk, but it can also remove the very flow that corrects mispricing.

“Manipulation” is also often misdiagnosed. A late move near a deadline can be someone with better information finally showing their hand, or it can be positioning into settlement. Thin liquidity makes both look the same on a chart. The right question is not “was it manipulated,” it is “could informed money enter, and was there enough depth for the price to mean anything.”

Regulation and platform differences in the US

In the U.S., the Commodity Futures Trading Commission is the key perimeter setter for event contracts. WealthManagement lays out a clean three-case map using Kalshi, PredictIt, and Polymarket.

Kalshi is the example of a prediction market operating inside the perimeter. WealthManagement describes Kalshi’s approval as a CFTC Designated Contract Market and frames it as exchange-style supervision, including market integrity, financial safeguards, disclosures, surveillance standards, and anti-manipulation protections. The point for users is not branding. It is that the venue is operating under a formal rulebook and oversight architecture.

PredictIt is the example of a narrow carve-out. WealthManagement notes PredictIt operated under a 2014 CFTC no-action letter with specific limits, including investment caps, participant limits per contract, and a research-oriented political forecasting scope. It also notes the CFTC withdrew that no-action letter in 2022, with litigation allowing continued operation during the court process. That history matters because it shows how tight the exemption box is and how quickly it can become contested.

Polymarket is the enforcement case. WealthManagement describes a 2022 CFTC enforcement action that resulted in a civil penalty and requirements to restrict U.S. access to markets under CFTC jurisdiction, followed by restructuring and continued operation in a more limited form. For anyone looking at prediction markets crypto, this is the reminder that “on-chain” does not automatically mean “outside U.S. rules” when the product functions like an event contract and is offered to U.S. users.

Practical takeaways for new participants

The first skill is reading signal quality, not just reading the number. Before treating a price as an implied probability, check whether the market can attract informed money. Fortune’s reporting on insider restrictions shows that venues may explicitly bar the very participants who would have the best information on certain contracts. That can be the right policy choice, but it changes what the price represents.

The second skill is separating contract design from narrative. A binary contract is easy to quote on social media, but it is still a tradable instrument with spreads and fees. A sudden jump can be late information, a liquidity vacuum, or traders leaning into settlement dynamics. If a move happens close to the deadline, the default assumption should be that new information or positioning arrived, then the sanity check is whether liquidity and participation were sufficient for the move to be meaningful.

The third skill is understanding the regulatory wrapper before worrying about edge. WealthManagement’s three reference cases draw a bright line for U.S. users: a CFTC-regulated venue like kalshi operates under exchange supervision, a no-action framework like PredictIt’s was narrow and contested, and a venue that served U.S. users without registration like polymarket faced enforcement and had to restrict access. That is not a moral judgment. It is a constraint that determines what you can trade and how reliably the venue can operate.

Finally, treat platform “accuracy rankings” as marketing until proven otherwise. Manifold’s own accuracy discussion shows how much the score depends on methodology and market conditions. If a platform is “more accurate” because it only lists easy-to-resolve questions or because it has deeper liquidity on a few headline markets, that does not generalize to every contract you might look at.

The Take

I’ve watched people treat a prediction market print like it’s a law of nature, then get surprised when the number was mostly a function of who was allowed to show up. Fortune’s 2026-04-26 reporting on Kalshi and Polymarket tightening insider rules is the clean example. If the venue bans politicians from trading their own races or employees from trading employer-linked contracts, the market may look more “legit” to regulators while also losing the sharpest information.

The expensive misconception is “price equals truth.” A prediction market price is a tradable clearing level for an event contract. It becomes a useful forecast when liquidity is real, informed participation is allowed, and the resolution process is credible. When any of those legs is missing, the implied probability is still a number you can quote, but it is not a number you should trust the same way.

Sources

• WealthManagement.com
• Fortune
• Manifold Markets
• Manifold Markets