Prediction Markets vs Polls

Polls and prediction markets both try to tell you what is going to happen. They measure completely different things, and knowing which one you are reading prevents most of the confusion around election coverage.

What each one measures

A poll is a sample of stated intentions at a point in time, with a margin of error and a weighting model applied by the pollster. A market price is what people will risk money on, aggregated continuously. A poll answers "who says they will do what"; a market answers "what do informed people expect the final result to be".

Markets consume polls

Traders read polls, turnout models, historical error and news, then price the outcome. So a market is not an alternative source of information — it is a weighted summary of every source, including the polls themselves, plus a judgement about how reliable they have been lately. That is why a market can move the moment a poll is published, and why it may barely move if the poll matches expectations.

Where markets beat polls

• Speed. Prices reprice within seconds of news; polls take days to field and publish.
• Aggregation. One price absorbs every poll, model and rumour at once.
• Skin in the game. Being confidently wrong costs money, which disciplines opinions.
• Non-polled questions. Nobody polls "will this bill pass committee by March" — a market can price it.

Where polls beat markets

• Detail. A poll tells you which demographics moved and why; a price tells you only the total.
• Thin markets. Low-volume political markets can be moved by a single motivated trader.
• Restricted crowds. When traders in the country in question are geoblocked, the market's local knowledge suffers.
• Long horizons. Markets a year or more out are dragged towards the middle by the cost of tying up capital.

Reading them together

Use polls for the mechanism and markets for the summary. If a poll shows a 4-point lead but the market is barely above 50%, the market is telling you it does not trust the polling, or it expects the position to change. That disagreement is usually the most informative thing on the page.

A caution about certainty

Both tools are probabilistic. A 15% candidate winning is not a failure of the market any more than a poll inside its margin of error is a failure of the pollster. Judge either only over a long run of events.

The political prices shown across this site come from the platform named on each card. They are market-implied, not our own view of any election.