Explore Prediction Markets

Metaculus Explained: How the Forecasting Platform Works

Metaculus is a forecasting platform where people make and refine probability predictions about real-world questions—like elections, economic indicators, scientific milestones, and geopolitics—but unlike prediction markets, you are typically not trading for profit. Instead, Metaculus aggregates forecasts, tracks accuracy over time, and uses scoring systems and community processes to surface well-calibrated probabilities.

If you came here wondering whether Metaculus is “like Polymarket or Kalshi,” the key difference is simple: Metaculus is primarily a forecasting-and-reputation system, not an event-contract exchange. You submit predictions and build a forecasting track record rather than buying and selling contracts with money.

Metaculus, in plain English: what it is (and what it is not)

Metaculus is best understood as a structured forecasting community and research platform. It hosts questions with clear resolution criteria and deadlines, lets forecasters submit probabilities or numeric estimates, and then evaluates performance once outcomes are known.

Metaculus is not, in its core product, a regulated venue for trading event contracts. That matters if your search intent is “how do I deposit, trade, and withdraw.” With Metaculus, the main “returns” are informational (better estimates) and reputational (a public track record), not financial.

That said, Metaculus can still be highly relevant to prediction-market users because it produces crowd probabilities that many people use as an input alongside market prices, polling, models, and subject-matter research.

How Metaculus questions work: the building blocks of a forecast

Most Metaculus questions are written to be resolvable—meaning that by a specific date (or when a criterion is met) the platform can say “yes” or “no,” or record a numeric value (like an inflation print). Clear resolution standards are the heart of the system, because accuracy scoring depends on unambiguous outcomes.

Common question types include:

  • Binary questions (Yes or No), expressed as a probability from 0 percent to 100 percent.
  • Numeric questions, where you forecast a value (often with a distribution or percentiles).
  • Date or time-to-event questions, where you forecast when something will happen.

Each question typically includes a resolution source (for example, an official government release, a reputable dataset, or a specified publication) and a resolution procedure. If you are used to event contracts, this is analogous to settlement rules and an oracle source—just without a cash payout.

From individual predictions to “the community forecast”: how aggregation works

Metaculus collects individual forecasts and publishes an aggregated forecast. The exact aggregation method can vary by question type and platform iteration, but the idea is consistent: combine many forecasts in a way that tends to outperform the average individual.

You will often see multiple lines or forecast views, depending on the question:

  • A community-style aggregate (the main “crowd” view).
  • Variants that incorporate track-record weighting or other methods when available.
  • Historical forecast charts, showing how probabilities moved over time.

If you are comparing this to markets, think of Metaculus as producing a probability “signal” more like a continuously updated consensus model than a price discovered by trading. There is no order book, no spreads, and no concept of liquidity in the trading sense—though you can still talk about “participation” and “responsiveness” as a rough analogy.

For readers who want to connect forecasts to market intuition, ProbabilityWire’s guide to implied probability can help you translate between “percent chance” language and price-like thinking in event contracts.

Scoring, calibration, and reputation: why Metaculus feels competitive without money

Metaculus is built around measurable forecasting skill. After a question resolves, forecasters receive scores based on how accurate and well-calibrated their predictions were.

In practical terms:

  • You are rewarded for being accurate and penalized for being confidently wrong.
  • Updating your forecast as new information arrives can improve performance, because scoring generally reflects your final (or time-weighted) stated beliefs rather than a one-time guess.
  • Over time, your public (or profile-based) track record becomes a reputation asset.

This is one of Metaculus’s biggest advantages versus casual polling: it creates incentives to think in probabilities, document reasoning, and learn from outcomes.

What “tournaments” and special forecasting programs change

Metaculus also runs structured forecasting efforts that look more like competitions or research programs than an open forum. These can include curated question sets, deadlines, and sometimes different visibility or participation rules.

The practical takeaway is that not every question on Metaculus is “just a public prompt.” Some questions exist because an institution, research team, or program wants systematically generated probabilities and written rationales. If you are a serious forecaster, these programs can be a pathway to demonstrating skill in a more organized environment.

Forecasting isn’t trading: how Metaculus differs from prediction markets

People often lump Metaculus into “prediction markets,” but it is more accurate to call it a forecasting platform. Here is the difference that changes everything:

  • In prediction markets, probabilities are inferred from prices, and prices are shaped by capital, risk limits, fees, and liquidity.
  • On Metaculus, probabilities come directly from forecasters’ stated beliefs, shaped by scoring incentives and community norms rather than profit and loss.

This distinction matters when interpreting the numbers. A market price reflects tradable consensus under constraints (and sometimes manipulation attempts that cost money). A Metaculus forecast reflects a scoring-driven consensus that can be highly thoughtful, but not “stress-tested” by financial risk in the same way.

Metaculus vs Polymarket vs Kalshi: a practical comparison

Below is a reader-focused comparison that highlights the differences that usually matter most. Exact product details can change, and trading platforms frequently update fees, availability, and product scope, so always verify on the platform’s official documentation before acting.

  • Platform type: Metaculus is a forecasting and aggregation platform; Polymarket is a prediction market using event contracts; Kalshi is a regulated event contracts exchange in the United States.
  • Main user action: On Metaculus users submit probabilities or numeric forecasts and update reasoning; on Polymarket users buy and sell outcome shares (typically Yes or No); on Kalshi users buy and sell event contracts (Yes or No).
  • Money at risk: Metaculus generally has no cash trading in the core experience; Polymarket and Kalshi both involve funds in trading.
  • “Price” mechanism: Metaculus uses an aggregated forecast from submissions; Polymarket and Kalshi both use market prices shaped by trading and liquidity.
  • Output you use: Metaculus provides forecast probability, history, and rationales; Polymarket and Kalshi both provide tradable prices and implied probabilities.
  • Best for: Metaculus suits research-driven probability estimates and track record building; Polymarket suits market-based sentiment and tradable probabilities; Kalshi suits event contracts with a regulated structure in the United States.

Reading a Metaculus forecast without overreading it

A Metaculus probability is best treated as an aggregated estimate, not a promise. It can be extremely informative, especially when the question is well-specified and the forecaster pool is engaged, but it is still a human-generated signal.

  1. Read the resolution criteria
    Identify what counts as Yes or No, or which numeric outcome is being forecast. Check the named resolution source, such as an official release, dataset, or publication.
  2. Check definitions and timing
    Look for the precise metric definition, including seasonally adjusted versus unadjusted or headline versus core figures. Check the resolution date, whether preliminary or final data applies, and whether later revisions count.
  3. Assess how current the forecast is
    Review the forecast history and recent updates. Metaculus can move slowly on fast-breaking news: an active market may reprice before forecasters log in and revise their estimates.
  4. Consider participation and research quality
    Read the reasoning behind forecasts and consider how engaged the forecaster pool is. On niche, research-heavy questions where markets are thin or absent, a forecasting community may provide a more useful estimate.
  5. Compare with other signals
    Use market prices, polling, models, and subject-matter research as cross-checks. Account for liquidity and other trading constraints when comparing forecasts with market-implied probabilities, and treat neither as a certainty.

Resolution rules and disputes: what happens when reality is messy

Metaculus questions are designed to resolve cleanly, but real-world data can be delayed, revised, or contested. In those cases, the important thing is the question’s written resolution policy—the stated source, timestamps, and definitions.

The resolution checks above matter because an apparently straightforward question can depend on a particular data release or a precise definition. This is similar to settlement mechanics on event-contract platforms, except that the outcome affects scoring and historical records rather than cash settlement.

Why Metaculus can be valuable even if you only care about markets

Even if you primarily trade event contracts, Metaculus can be useful as a cross-check:

  • It can highlight when a market price looks “off” relative to a research-oriented crowd.
  • It provides written rationales and links that can improve your information set.
  • It offers a track-record lens: you can identify consistently strong forecasters and learn how they reason.

Some traders use Metaculus to generate priors (a starting probability) and then adjust based on market microstructure, fees, and timing considerations.

What Metaculus does not provide (and what to use instead)

If your goal is specifically to “take a position” with money, Metaculus is not a substitute for an event-contract venue. It does not function like a brokerage, and it is not designed around deposits, withdrawals, minimum trade sizes, or order execution.

For readers looking for tradable products, you will usually want to compare venues like Kalshi (event contracts), Polymarket (crypto-based prediction markets), or brokerage-integrated alternatives when available.

Metaculus explained in quick answers

Not in the typical trade-contracts-for-profit sense. Metaculus is primarily a forecasting platform where people submit and update probability estimates, then get scored on accuracy.

Metaculus’s core experience is not designed as a cash trading platform. Any financial incentives, if offered through special programs or external arrangements, depend on the specific program rules and should be verified in the program’s official terms.

Markets reflect tradable prices shaped by liquidity, fees, participant constraints, and risk. Metaculus reflects an aggregated belief from forecasters responding to accuracy-scoring incentives, not trading profit and loss.

Each question includes resolution criteria and a specified source, such as an official dataset or a named publication. The question resolves based on those written rules, which is why reading the fine print matters.

It can be, especially for well-defined questions with clear data sources. Still, treat the probabilities as informed estimates, not certainties, and compare them with other signals, including market-implied probabilities and official economic releases.