Andrews wins South Carolina as markets capture broader Democratic sweep

Democrat Annie Andrews defeated Republican Darline Graham in South Carolina, delivering the night’s first prediction-market upset, according to Reuters’ November 4 wrap of the midterm results.

Graham had been favored to win, with market odds of about 70 percent. Her defeat marked a notable exception to what Reuters described as a broadly correct market outlook for the election.

A surprise within a wider Democratic victory

Democrats took control of both the Senate and the House, according to Reuters. Prediction markets largely anticipated that overall outcome, even as the South Carolina result went against the favored candidate.

The distinction matters when assessing forecasts: correctly anticipating control of Congress does not mean every individual race will follow market expectations. A candidate priced at about 70 percent remains a favorite, not a certainty.

Platforms point to accuracy assessments

Kalshi said its own accuracy assessments found it more accurate than other leading forecasting models, Reuters reported. Separately, a person familiar with the matter told Reuters that Polymarket’s election forecasts were about 98 percent accurate.

The supplied summary does not specify the methods or coverage behind those assessments. The figures therefore do not establish a directly comparable measure of performance between the platforms.

What it means for prediction markets

• Favorites can lose: Graham’s roughly 70 percent market odds left room for an Andrews victory; the upset alone does not establish that the probability was poorly calibrated.
• Overall direction and individual races are distinct: Markets broadly anticipated Democratic control of both chambers while missing the favored outcome in South Carolina.
• Accuracy claims require context: Kalshi’s self-assessment and the reported Polymarket figure have different attribution, with no methodology provided in the summary.