Can you trust the price?

Here is the objection that stops most newcomers cold. A prediction market is open to anyone with a balance, and the price is just whatever people are willing to pay. So what stops a rich, motivated actor from simply buying the answer they want — bidding a contract up to 80¢ to make the world believe an event is likely, when really it's a coin flip? If the number is that easy to push, why would anyone trust it?

It's the right question. And the answer is the most elegant thing about the whole machine: a price pushed off the truth is free money for everyone else. The same openness that lets a manipulator in lets a hundred other people take the other side of his trade — and they're motivated by exactly the thing that motivates him, profit, except the truth is on their side. Let's take the three worries one at a time: manipulation, insider trading, and thin markets.

Worry one — painting the tape

Start with the blunt-force version: a whale walks in and buys, buys, buys, dragging a contract from 50¢ to 80¢ to manufacture a headline. Traders call this painting the tape — moving a price not to express a view but to plant a false impression. Can it be done? For a moment, yes. Any market can be shoved if you spend enough. The question is never whether you can move the price; it's whether you can keep it there.

You can't, and the reason is arithmetic. When you buy a contract up to 80¢ that's truly worth 50¢, you are posting a public, standing offer: I will pay 80 for a thing that pays out 50. Every informed trader who sees that sells into you. Arbitrageurs who do nothing all day but hunt mispricings fade your move. They keep selling — and their selling drags the price back down — until it's back at fair value and your money is in their pockets. Your manipulation didn't move the world's belief. It subsidized your opponents.¹

This isn't a hopeful theory; it's been tested on purpose. In controlled experiments where some traders were explicitly paid to push the price one way, researchers found the manipulators essentially failed: other participants leaned against them and prices stayed close to where the fundamentals said they should be. Manipulation transferred money to the rest of the market without durably distorting the forecast.²

And it has happened with real money on the line. In the 2012 US presidential race, a single trader on the prediction site Intrade spent something on the order of a few million dollars propping up the Romney contract — and each time he pushed it up, other traders sold it back down, the price snapped toward consensus, and the campaign ended; the spending bought a temporary bump and a large loss, not a changed outcome.³ That is the rule, not the exception: manipulation in a liquid market is expensive and usually unprofitable — you pay to be wrong in public.

The intuition that "money can buy the price" has it exactly backwards. In a market with depth, money is what punishes a false price. The more capital a manipulator commits, the larger the gift he's handing to the people on the other side. Pushing a price is not exercising power over the market — it's announcing a sale and waiting for buyers.

Worry two — the insider

Now the second fear, and the one with the genuinely surprising answer. What about someone trading on private information — the staffer who knows the announcement before it's public, the pollster sitting on numbers no one else has seen? In the stock market we have a name for this and a law against it: insider trading, and it's a crime. Surely it poisons a prediction market too?

Here the two kinds of market come apart, and you have to be precise about why insider trading is bad in equities. When an insider trades a stock, the harm is distributional: they're transacting against a counterparty who doesn't know what they know, and they take that person's money. The information eventually surfaces anyway — the insider mostly just front-runs it. The damage is the theft, not a lie in the price.

A prediction market exists to do one job: produce an accurate forecast. Measured against that job, the person with private information is not a parasite — he's the sensor. When the staffer who knows the real number buys, the price moves toward the truth, and it moves before the public announcement, which is precisely the point of a forecasting instrument. The "insider" isn't stealing the signal; he's contributing it. His trade is how the market learns something the crowd didn't know.⁴

This is not a loophole to be embarrassed about; it's part of why these markets forecast so well. They pay people to go find out what's true and put it into the price. Someone with an edge — a private poll, a source, a model no one else has run — is rewarded for revealing it through their trade, and the forecast gets sharper for everyone reading it. A market that perfectly excluded everyone who knew anything would be a market that knew nothing.⁵

Worry three — the thin market

So far the news is good — too good to leave there, because both arguments quietly lean on the same word: liquid. Manipulation fails because informed traders and arbitrageurs are standing by to fade it. Private information gets priced in because there's a real market for it to move. Take the depth away and both guarantees evaporate.

This is the worry that's actually warranted. A thin, illiquid market is noisy and genuinely manipulable. If a question has a few thousand dollars of open interest and trades twice a day, there's no standing army of arbitrageurs watching it. A single motivated buyer can move it and there may be no one on the other side to push back — and even absent any manipulation, the last trade might be stale, a leftover from a bored gambler, not a considered estimate. A 5,000-dollar market on some obscure question is not an oracle. It's a rumor with a number attached.

So the honest rule isn't "trust the price" or "don't" — it's trust the price in proportion to the market's depth and volume. A contract with millions in daily turnover, hundreds of participants, and arbitrageurs watching it is a serious estimate that fights to stay honest. A near-empty market is a guess. Same mechanism, wildly different reliability — and the difference is liquidity. Before you believe a number, look at the volume underneath it.

What actually keeps it honest

Notice that none of these defenses is a rule, a regulator, or a referee. No one bans the manipulator from trying or the insider from trading. The market polices itself, and the police are other traders with money on the line. That's the thread running through this whole series: self-correction runs on skin in the game. A false price is a profit opportunity, and a liquid market is a crowd competing to claim it — which is exactly why, as I argued in "A price is a probability," these markets end up so well-calibrated. The same force that grades the forecast is the one that defends it.

A fair caveat for this series: most of the money in these markets today is on sports — roughly 80% of volume on the largest US venue — so a lot of what keeps prices honest is people trading for the thrill, not forecasters chasing truth. But the mechanism doesn't care about the motive. A sharp who's fading a manipulator to make a buck is defending the price just as surely as a wonk correcting a poll; greed and accuracy point the same way. That's the quiet genius of it — the market doesn't need its participants to be virtuous, only to want to win.

So: can you trust the price? Not on faith, and not blindly. Trust it because a price pushed off the truth is a bounty that the whole market is racing to collect — and trust it more the deeper and busier the market is. A manipulator can rent a wrong number for an afternoon; he can't own it. The truth, in a liquid market, is the cheapest thing to bet on — which is why, over and over, it's the thing the price comes home to.