Prediction Markets & Geopolitical Risk
How prediction markets work — the science behind the numbers, and where they fall short
In 1907, Francis Galton watched 800 people at a county fair guess the weight of an ox. The median guess — 1,207 pounds — was within nine pounds of the actual answer, beating nearly every individual in the crowd, including the livestock experts. The phenomenon he stumbled on, now called the wisdom of crowds, turns out to be surprisingly durable: when people form estimates independently and have real skin in the game, their errors tend to cancel out in ways that beat most individual experts.
Prediction markets are built around exactly this mechanic. Traders buy and sell contracts on specific outcomes — if the event resolves YES, the contract pays out — so anyone who thinks the current price is wrong has a direct financial reason to say so. That incentive structure is what separates markets from polls or punditry.
Philip Tetlock's research on forecasting reinforced why this works. His superforecaster studies found that careful generalists — people who track their predictions, update often, and think in probabilities — consistently outperform government intelligence analysts on geopolitical questions. Aggregating a pool of such forecasters improves accuracy further still.
On everyday political and economic events, the track record is strong. Well-traded markets are well-calibrated: events priced at 70% happen roughly 70% of the time over large samples. They tend to beat election polls and react to breaking news faster than most institutional forecasters.
The harder problem is rare, sudden crises. A war might sit at 4% for years before it happens — and that's not a failure, it's what a 4% event looks like when it occurs. For low-probability, high-stakes scenarios like this one, treat these numbers as a continuously updated risk gauge, not an early-warning system.
Volume matters too. A market with $500k in total trading is much harder to distort than one with $3,000. Thin markets can be moved by a single trader and deserve extra scepticism — this site filters out the thinnest ones, but check the volume figure on each card.
The platforms
The largest prediction market by volume. Uses USDC (a crypto stablecoin) and operates offshore without formal regulation. Its geopolitical markets have the highest liquidity and tend to set the reference price that other platforms follow.
The first CFTC-regulated prediction exchange in the US, trading in real dollars. Smaller volume than Polymarket on most geopolitical questions but growing. The regulatory structure makes it more straightforward for US participants.
A free-to-use platform trading in virtual "mana" with no monetary value. Without financial incentives, estimates rely on intellectual honesty and tracked calibration scores. Covers a much wider range of niche geopolitical questions than real-money platforms will touch.
A structured forecasting platform where users submit probability estimates and are scored on calibration over time. No money changes hands — accuracy is tracked through reputation scores. Draws a strong community of analysts and careful forecasters; well-suited to longer-horizon questions.
How to read the numbers on this site
- Probability % — the current consensus estimate. 12% means roughly 1-in-8 odds at this moment, across the traders or forecasters on that platform.
- Volume — total USD traded (real-money markets only). Higher volume = harder to distort. Be sceptical of anything under ~$5k.
- Liquidity — capital immediately available to trade against. Low liquidity means prices are easier to move with a single order.
- Composite score on the dashboard is a simple average across tracked markets — a rough directional barometer, not a precise forecast.
Frequently asked questions
What is a prediction market?
A platform where people trade contracts on real-world outcomes. If you think something will happen, you buy a YES contract. If it resolves YES, the contract pays out — if it doesn't, you lose your stake. The going price (e.g. 0.30) implies a roughly 30% probability. Because money is at stake, participants have a direct incentive to estimate honestly.
Is this just 'wisdom of the crowds'?
It draws on that principle, but with an important addition: an incentive structure. The classic wisdom-of-crowds effect (groups averaging out individual errors) works best when people have good information and genuine motivation to be accurate. Real-money markets provide that motivation. Play-money and reputation-based platforms like Manifold and Metaculus rely more on intellectual honesty and tracked calibration scores.
How accurate are these forecasts?
On political and macroeconomic events, well-traded markets are consistently well-calibrated: events priced at 70% happen roughly 70% of the time over large samples. They tend to outperform polls and expert panels on election outcomes and react to new information faster than most institutional forecasters. The weak spot is rare, high-stakes events like wars — too infrequent to build a solid calibration record. Treat those numbers as directional signals, not precise odds.
What is Polymarket?
The largest prediction market platform by volume, using USDC (a crypto stablecoin). It operates offshore and is unregulated, but has the highest liquidity on geopolitical questions and tends to set the reference price that other platforms follow.
What is Kalshi?
A US-regulated prediction exchange (CFTC), using real US dollars. Smaller volume than Polymarket on most geopolitical questions, but growing quickly. Being regulated makes it more accessible and straightforward for US participants.
What is Manifold Markets?
A play-money forecasting platform. Users trade with 'mana', a virtual currency with no monetary value, which means anyone can participate for free. Without financial stakes, prices rely more on intellectual honesty than financial incentive — but Manifold covers a much wider range of niche questions than real-money platforms will touch.
What is Metaculus?
A structured forecasting platform where users submit probability estimates and are tracked on calibration over time. No money changes hands — accuracy is rewarded with reputation scores. The community includes domain experts and professional analysts, making it particularly useful for longer-horizon questions.
Can collective forecasting predict wars?
It can reflect aggregated human judgment about war risk, but rare, sudden events are inherently hard to price. Markets can sit at 5% for years, then an invasion happens — that's not a market 'failure' in statistical terms, it's what a 5% probability event looks like. The probabilities are best treated as a continuously updated risk gauge, not a reliable early-warning system.
How should I read the percentages on this site?
A market showing 12% means traders collectively price that outcome at roughly 1-in-8 odds at this moment. Volume and liquidity matter: a market with $500k traded is far more meaningful than one with $2,000. Very thin markets can be moved by a single trade and should be treated sceptically.