Virginia prediction markets gave McGuire an edge and Cline a commanding lead

A WSLS 10 segment on the eve of Election Day used prediction-market prices to portray sharply different levels of confidence in two Virginia congressional contests. Kalshi put Democrats’ chance of taking the U.S. House at more than 90%, while assigning Republican Rep. John McGuire a 64% chance of defeating former Democratic Rep. Tom Perriello in Virginia’s 5th District and Republican Rep. Ben Cline a 93.4% chance of beating Democratic challenger Beth Macy in the 6th.

The station reported more than $1 million wagered on the 5th District contest and more than $219,000 on the 6th. It described Kalshi and Polymarket as platforms on which users can put money on election outcomes nationwide.

As this publication reported Nov. 1, national polls and prediction markets had both signalled a strong Democratic advantage in the fight for House control, though the overlap did not necessarily amount to independent confirmation.

Trading interest in the 5th District had already stood out. A WJLA and WSET report in October called it Virginia’s most actively traded congressional race on Kalshi and Polymarket, with more than $839,000 traded across the two platforms, compared with about $200,000 in the 6th District and $90,000 in the 9th. David Richards attributed the heavier activity to uncertainty over the result, saying the 5th was a contest in which nobody was entirely certain of the outcome.

That uncertainty was reflected in competing indicators. WJLA and WSET said traders then gave McGuire only a narrow edge. By Nov. 2, Prediction.com recorded Polymarket pricing McGuire’s chances at 66%, up from 58% three days earlier. The figure was an implied probability of victory, not a forecast of his vote share.

A Washington Post-Schar School poll conducted Sept. 23-28 put McGuire ahead 51% to 45%, while an August Expedition Strategies poll found a 47% to 44% McGuire lead. Sabato’s Crystal Ball moved the race from Leans Republican to Toss-Up after the Washington Post-Schar poll, whereas Cook Political Report continued to list it as Likely Republican. State Navigate projected a Perriello victory by about two points on Oct. 29.

WSLS quoted Chapman Rackaway as saying prediction markets have roughly similar reliability to polls and are strongly connected to them. He argued that participants have money at stake and seek information before betting, and said markets can register reactions to scandals or new advertising more quickly than polls, which may need three to four days of data collection even for a quick survey.

Rackaway also cautioned that local races are harder to predict when fewer people participate, because a thinner betting pool creates a smaller sample. His own model gave Democrats a strong chance of taking the House and a 55% chance of taking the Senate, after he had viewed Senate control as evenly balanced until late in the previous week.

The broader rise of election contracts has prompted questions about what their prices measure. CNBC reported that congressional-election contracts had attracted more than three times the wagers placed on the 2024 races, but noted that gambling-platform odds use methods different from polling and are not a substitute for it. A Brennan Center for Justice-funded study found that voters were 10 to 20 percentage points more likely to believe a candidate would win when prediction markets favoured that candidate, regardless of polling.

CBC reported that more than $1.4 billion had been traded on U.S. midterm prediction markets by Oct. 23. Its reporting, citing the Anti-Corruption Data Collective, found many of the 21,811 markets examined were thinly traded, allowing relatively small wagers to move public-facing odds and potentially create an impression of changed electoral strength.