Updated
Updated · The Washington Post · Jun 23
Post Analysis Finds 75% Election Favorites Usually Win Across 268 Prediction-Market Cases
Updated
Updated · The Washington Post · Jun 23

Post Analysis Finds 75% Election Favorites Usually Win Across 268 Prediction-Market Cases

3 articles · Updated · The Washington Post · Jun 23

Summary

  • A Post review of 268 candidates priced at 70% to 80% on Kalshi or Polymarket found roughly three in four won, suggesting recent headline misses did not break the markets’ overall accuracy.
  • That result helps explain losses by market favorites such as Spencer Pratt in Los Angeles, where Kalshi’s mayoral market topped $40 million in bets before he finished third.
  • Across broader probability bands, calibration also held: candidates near 95% won almost always, while those near 50% won about half the time in the Post’s analysis of 2026 primaries.
  • Experts said bettors often misread probabilities as vote share, giving numbers like 75% a false sense of certainty even though unlikely outcomes still happen regularly.
  • Researchers and forecasters said prediction markets aggregate polls, models and trader judgment well, but still cannot replace polling for explaining voter behavior or why an electorate moves.

Insights

Are prediction markets the future of forecasting or just high-stakes gambling for regulators to solve?
As AI traders dominate, is the 'wisdom of the crowds' being replaced by algorithmic supremacy?
Can markets on geopolitical events exist without threatening national security through insider trading?