Governments set policy on forecasts. Newsrooms lead with them. Companies commit capital to them. Families decide where to live, study, and invest because of them. Almost all of it rests on two foundations — opinion polls and expert commentary — and both share the same flaw: nobody pays a price for being wrong.
Talk is free
A poll asks what you think, at no cost to you. There is no penalty for answering carelessly and no reward for answering well, so people guess, they satisfice, and they say what sounds respectable. Response rates have collapsed, which means the sample is increasingly made of whoever still picks up the phone. And a poll is a photograph of the present, taken weeks ago, being used as a forecast of the future.
Experts are not much better
Philip Tetlock spent two decades scoring the concrete predictions of political experts against what actually happened. His finding was uncomfortable: the average specialist performed close to chance, and the most confident, most quoted commentators tended to score worse than their quieter colleagues, not better. Being wrong loudly costs a pundit nothing. There is no ledger, and next year the same voices are booked again.
What changes when money is on the line
A market price is a forecast someone paid to make. If you believe a price is wrong, you can take the other side — and if you are right, you are paid for it. That single mechanism does three things no survey can.
It filters: people who do not really believe what they are saying tend not to fund it. It aggregates: Friedrich Hayek's insight was that the knowledge a society needs is never held by any one authority — it is scattered in fragments among people who each know something small and specific, and a price is the instrument that pulls those fragments together. And it updates: continuously, in public, the moment anyone learns anything, rather than whenever the next survey is commissioned.
Why this matters most here
In Greater China, the distance between what people say and what they believe is not a rounding error. It is the entire problem.
The economist Timur Kuran named this preference falsification: under social or political pressure, people misrepresent their real views, and they do it so consistently that the public record stops describing reality. Ask directly about leadership, cross-strait risk, or the true state of the economy, and many people will give you the safe answer, the useless answer, or no answer at all. Not because they are dishonest — because answering honestly can be costly.
Polling fails hardest exactly where the questions matter most.
A market asks nothing of you. You never state an opinion, make a claim, or attach your name to a view. You take a position — anonymously, with your own money. What emerges is a number that no one had to be brave to produce.
Who this is for
A market price is not a headline. It is an instrument, and it is most useful to people who have to act on the future rather than describe it.
- Governments and policy staff. A live estimate of how a decision is being read, from participants who are not briefing you and have no incentive to flatter.
- Journalists and editors. A citable, timestamped number with a public methodology — and a record that can be checked afterwards, unlike a quote from an unnamed analyst.
- Boards and management teams. Political and economic risk is the hardest input in any strategic decision, and the one most often outsourced to a consultant's slide. Whether to expand into a market, sign a long-term supply contract, hedge a currency, delay a listing, or move production — each turns on a question about the future that someone has to price. Here that price is public, live, and updated by people with money at stake rather than a fee to collect.
- Investors. A probability you can compare against what is already priced into your own book.
- Researchers. A continuous, timestamped series, rather than sparse survey waves with shifting methodology.
- Families. Where to send a child to school, whether to move, whether it is safe to go home. The decisions that made us build this.
What a price actually means
When a market reads 41¢, it is saying the crowd — weighted by how much each participant is willing to risk — puts roughly a 41% chance on that outcome. It is a probability, not a prophecy. Things priced at 41% are supposed to happen about 41% of the time, and a market that never surprises you is a market that is not telling you anything.
Where markets fall short
We would rather you knew the limits than discovered them later. Thin markets are noisy: a price with little money behind it carries little information. Markets can be confidently wrong, particularly about rare events and about things people badly want to be true. Manipulation is a genuine risk, which is why our market integrity rules exist and why every market must name its resolution source in advance. And some questions simply cannot be settled objectively, so we do not list them.
We publish these limits deliberately. A signal you cannot interrogate is not a signal — it is just another opinion, and the world has plenty of those.
真策 · An index of truth for Greater China.