Key Finding

Prediction markets are designed to price a headline once and then hold that price. News breaks, traders absorb the information, the market adjusts, and the new price remains. If this were the full picture, the initial move following a headline would be the definitive move, and maintaining it would be the norm.

It is not. A study by Vera Research, analyzing more than 60,000 reactions of two cents or greater on Polymarket — one of the largest blockchain-based prediction markets, where users trade on the outcomes of real-world events — between April 29 and June 25, 2026, found that these reactions fall into just five recurring trajectory shapes over the four hours following each headline. The clean move that ramps and holds — the shape most people assume describes every reaction — accounts for barely one in three.

The remaining four shapes behave in ways that the conventional model does not account for: they grind upward late, overshoot past their first peak, fade back toward zero, or reverse course entirely and cross to the opposite side of where they began.

The One in Three Moves Is Fake study · Vera Research

Three Shapes That Hold or Extend the Move

Three of the five trajectories ultimately hold or extend the initial reaction:

• Snap-and-hold (29.1%) ramps toward its peak within the first hour and maintains that level.
• Slow-grind (25.3%) barely moves in the early period, then reprices later.
• Accelerating (16.2%) never settles — by hour four, it sits well beyond its early peak.

Together, these three account for 70.6% of all measured moves, representing reactions that end up meaning what they initially appeared to mean.

Two Shapes That Round-Trip

The remaining two trajectories give back the move entirely:

• Spike-and-fade (21.7%) pops quickly, then slides back to approximately zero by hour four — the same fade pattern equity traders observe after an earnings pop.
• Reversal (7.7%) does not simply give back the move. It ends hour four on the opposite side of where it began.

Combined, these two shapes represent 29.4% of everything measured. Nearly one in three moves hands itself back.

Independent Verification Through Raw Counting

The round-trip share does not depend on the five-cluster taxonomy. Direct counting tells the same story. Of every 100 moves measured, 27 keep less than half of their peak by hour four, and 22.4% finish flipped against their own peak, ending on the far side of where they started. The median move retains just 0.750 of its peak.

Two independent re-derivations of the five groups, run from scratch with a different random seed, matched the original classification at 0.985 out of 1.00, confirming that the taxonomy is not an artifact of a single run.

News Category Does Not Predict the Shape

A natural question is which types of news produce which shapes. The answer, according to the data, is almost none of them. Every news category Vera tracks produces all five shapes in broadly similar proportions. Geopolitics does not spike more than macro; economics does not grind more than politics. The folklore that "geopolitics spikes and macro grinds" does not survive measurement.

The implication is straightforward: knowing a headline's topic tells you little about whether its resulting move will hold or reverse. The first price observed after a headline is often the least reliable piece of information available, and the news category itself provides no warning about which kind of move is unfolding. The shape is what carries the signal, and the shape only reveals itself over time. This stands in tension with the growing practice of citing prediction market prices as real-time indicators of event probabilities, since a substantial share of those prices do not stabilize until well after the initial reaction.

Summary of Findings

More than 60,000 measured Polymarket reactions (every move of two cents or more, April 29 to June 25, 2026) sort into five recurring shapes over the four hours after each headline:

• Three shapes hold or extend the move: snap-and-hold (29.1%), slow-grind (25.3%), and accelerating (16.2%) — together 70.6% of moves.
• Two shapes round-trip: spike-and-fade (21.7%) and reversal (7.7%) — together 29.4% of moves, nearly one in three.
• Counted directly without clustering: 27% of moves keep less than half their peak by hour four, 22.4% finish flipped against it, and the median move keeps 0.750 of its peak.
• The five groups reproduce from scratch under an independent audit with a different random seed, matching the original at 0.985 out of 1.00.
• News category does not predict the shape: every category produces all five shapes in broadly similar proportions.

Method and Data

The study is built on the Vera dataset: more than 60,000 measured Polymarket reactions of two cents or more, spanning April 29 to June 25, 2026, each classified by the trajectory shape it traced over the four hours after its headline. Percentage shares are derived from a five-cluster analysis of peak-normalized trajectories, re-derived from scratch by an independent audit at 0.985 out of 1.00.

Odds referenced are each market's own price at the time Vera flagged the story, stated as fact rather than advice. Nothing in the study constitutes a recommendation to buy, sell, or trade any market or asset. Vera and Crypto Briefing are not registered investment advisers.