Prediction Markets ’Attention Gap’ Study: News Repricing Doesn’t Match the Biggest Wires
A Vera Research study using Polymarket data finds a “Attention Gap” in prediction markets: prices reprice based on where market attention already sits, not on newsroom pecking order.
Across 56 days ending 25 June 2026, the dataset covered 618 news sources matched to 3,434 Polymarket markets, with 476,000 scored story-market pairs. The all-source baseline showed a 15.2% chance that prices moved by at least one cent within the next hour (the 1.0x line).
Top financial wires underperformed this baseline on the same measure: Bloomberg (0.70x), Reuters (0.74x), Wall Street Journal (0.75x), and Financial Times (0.90x). Meanwhile, three niche specialists outperformed and clustered higher between 1.5x–1.6x: @financialjuice (1.55x), @IranIntl_En (1.55x), and @solidintel_x (1.60x).
The report tested controls for market-category bias and for using the same markets. Category mix explained about 51% of the gap, but specialist accounts still ranked above wires after controls. The same-market control left roughly 15% of the separation and broke the original ranking.
Key trading implication: prediction markets can move most when attention is already building (busy-hour clustering) rather than when major outlets publish first. This can affect timing strategies and news-driven positioning, but the study measures past behavior, not a forward forecast.
Neutral
The study suggests the effect of news on prediction markets is highly dependent on attention already present in the market. That reduces the reliability of “big wire = immediate price move” assumptions, which is a subtle drag on simple headline-following strategies. However, it also provides a framework traders can exploit: monitor when markets are already active (busy-hour clustering), not only which outlet published.
Short-term, this can lead to fewer predictable, one-to-one reactions to major headlines and more sensitivity to timing and audience concentration. Traders may adjust by watching order-book/volume proxies for attention rather than treating publication as the trigger.
Long-term, the results align with past patterns in markets where liquidity and participation determine responsiveness to information (similar to how in traditional markets, higher participation often amplifies informational moves). Still, the report’s metric is retrospective (1-cent move within an hour) and doesn’t identify whether the mechanism is “attention clustering” or “story already priced in,” so conviction should remain moderate. Overall, market stability impact is likely limited; the main change is strategy quality rather than market fundamentals.