Midterm bets exceed $1.48 billion as forecasting accuracy faces scrutiny

More than $1.48 billion has been wagered across prediction markets on Tuesday’s US midterm elections, putting a substantial pool of election forecasts to the test as results arrive.

Kalshi reported $1.33 billion in midterm trading volume on Monday, according to the supplied summary. About $145 million has been wagered on Polymarket’s congressional markets.

A broad test of election probabilities

More than 28,000 midterm markets have opened. That breadth makes the forecasting question larger than whether traders correctly identify the outcome of any single contest: it also concerns how accurately probabilities reflect outcomes across many markets.

A 2025 Vanderbilt University study examining nearly 2,500 markets tied to the 2024 election found that prediction markets are generally well calibrated at scale. That finding provides context for assessing this year’s forecasts, but does not establish how accurately the midterm markets will perform.

What it means for prediction markets

• Volume and accuracy are separate measures. The more than $1.48 billion wagered establishes the scale of activity, not the reliability of the forecasts.
• The assessment extends beyond individual winners. With more than 28,000 midterm markets opened, aggregate calibration is a distinct question from whether a particular market’s favored outcome occurs.
• Earlier research offers a benchmark, not a verdict. The Vanderbilt University findings concern the 2024 election; the current midterm forecasts remain a separate test.