News Trading in Prediction Markets: Speed Is Not the Same as Being Right

News trading is the business of buying or selling within seconds or minutes of a headline. In prediction markets the first mover often gets the best price and the worst information. The edge is not raw speed. It is knowing whether the headline is true, whether it is already in the price, and whether it actually triggers the market's resolution rule.

Those are three separate questions, and a trade that fails any one of them is a bad trade even if the headline turns out to be correct.

Three filters, in order

Is it true? A headline is a claim, not a fact. The claim needs a primary source: the court docket, the official statement, the data release, the referee's decision. "Sources say" is a lower grade of evidence than "the filing says," and a screenshot is not a source at all.

Is it new? A true story can be old. Markets reprice on the first publication, not on the tenth retweet. If the price moved before you saw the item, you are not trading news; you are trading the residual.

Does it resolve this market? This is the filter most traders skip. Prediction markets resolve on a specific criterion, and the criterion is often narrower than the headline. "Reported" is not "confirmed." "Charged" is not "convicted." "Announces intention to resign" is not "resigns." A story can be entirely accurate and still leave the contract at its previous price.

Speed and the structure of the market

Polymarket's order book is hybrid: orders are matched offchain by an operator and settled onchain. All orders are technically limit orders, and a "market order" is a limit order priced to execute immediately against resting orders. That design gives fast matching, but speed is only worth something if the counterparty is slow.

In a heavily watched market, the first repricing often happens within seconds, and the remaining edge is not in the headline itself but in its interpretation. In a niche market, the price may not move for hours because nobody is watching. The second case is where a fast retail trader can still compete, and it is also where verification is hardest, because there is less coverage to confirm against.

There is a subtler cost: adverse selection. If you buy immediately after a headline and the seller is a professional who saw the same item first, the fill is evidence that you are late. The best-looking fills in a news spike are often the worst trades.

False reports and the cost of being early

False headlines are a recurring feature of fast news. Accounts get hacked, translations drift, parody accounts get screenshotted, old stories recirculate without dates, and "sources" turn out to be one person repeating another.

Consider a hypothetical. A market trades at 0.40. A screenshot claims an official has been indicted, and the price jumps to 0.75. You buy at 0.72. Four minutes later the claim is denied and the price returns to 0.42. You lose 30 cents per share, and the exit is a race against everyone else who made the same mistake.

The arithmetic of that loss is the argument for verification. If you size at $500, the error costs $200 or more. A two-minute check of the primary source costs nothing. The asymmetry is overwhelming, which is why the discipline is to accept missing some real moves rather than take unverified ones.

The resolution gap

Even a true headline can fail the third filter. Polymarket's documentation is explicit that the rules, not the title, define resolution, and that clarifications can be issued after trading begins.

A useful way to see the gap is to rewrite the headline as the contract sees it. "Prosecutors are preparing an indictment" becomes "Has an indictment been unsealed or announced by the named authority before the deadline?" Those are different facts. The first can be true while the second is false, and the price should not move to 0.95 on the first.

Before trading any news event, write down the exact sentence from the rules that would make the contract resolve Yes. If the headline does not match that sentence, the trade is not a resolution trade. It might still be a sentiment trade, but it should be sized as one.

How to trade news without guessing

1. Pre-commit to sources. For each market you follow, list the one primary source that will settle it. Follow that source directly, not an aggregator.
2. Use a two-source rule for size. One credible report is enough to watch; two independent reports are enough to trade small; the primary document is enough to trade full size.
3. Timestamp everything. Record when the headline appeared and when the market moved. If you cannot establish the order, you cannot learn from the trade.
4. Trade the second move. After the initial spike, markets often overshoot or undershoot. A slower trader who has verified the facts can trade the correction with better information than the trader who bought the spike.
5. Set a verification deadline. If you cannot confirm within a stated window, do not trade. Missing a move is a cost you can budget; a false-report loss is not.

The economics of a news scalp

At 0.50, a politics taker fee is one cent per share per side, so a one-cent scalp is zero expected value before spread. A three-cent move is barely a trade. A news scalp needs a move large enough to cover the round trip, the spread, and the probability that the report is wrong.

Suppose you estimate a 20% chance the report is false and the price fully reverses, a 60% chance it is true and moves the price 8 cents in your favor, and a 20% chance it is true but already priced. The expected gross move is 0.60 × 8 minus 0.20 × 8, or 3.2 cents. Round-trip fees and spread at 0.50 are roughly 4 to 6 cents. The trade is negative before any slippage. Most news scalps that feel obvious are not profitable after this arithmetic.

What the data cannot tell you

There is no public dataset that gives a general base rate for false headlines in prediction markets. The frequency depends on the event, the sources covering it, and the platform's resolution rules. Any claim that "most headlines are wrong" or "most are right" is a guess unless it is measured for a specific market and period.

Price data alone also cannot tell you causality. A jump after a headline is consistent with the headline causing the move, with a simultaneous trade by someone who saw the same thing, or with a liquidity gap. Without timestamps and order-book data, the story is a narrative, not evidence.

The discipline

The efficient-markets literature is a useful corrective here: public information tends to be incorporated quickly, and the semi-strong form of the hypothesis says prices reflect it. Prediction markets are not perfectly efficient, but the deviations are small and concentrated in places where attention is thin. That is not a licence to trade every headline; it is a reason to trade only the ones where you have a verifiable edge in interpretation or speed.

The traders who last at this are not the fastest. They are the ones who can answer "is it true, is it new, and does it resolve the contract" before they click buy, and who are willing to do nothing when the answer is unclear.