On the ballot
November 3

Each race starts where forecasters rate it today, averaged: the Senate from 11 forecasters (Cook, DDHQ, Econ, FPO, Fox, IE, NBC, RCP, Sabato, Silver, ST, Oct 20, 2026 to Oct 29, 2026), the House from 10 (Cook, IE, Sabato, Econ., ST, DDHQ, FPO, Fox, Silver, RCP, Oct. 22, 2026 to Oct. 29, 2026), as listed on Wikipedia. A seat every forecaster calls safe is placed by its lean, the 2026 Cook Partisan Voting Index. The slider is the national vote for the House, Democrats minus Republicans: "Polls today" is the average of 5 polling averages of the generic ballot (Decision Desk HQ, FiftyPlusOne, NBC News, RealClearPolitics, Silver Bulletin), October 27, 2026. Past years are the national House vote of 2008, 2010, 2018, 2022, 2024. Moving the slider moves every race by the same number of points; the Senate and House marks on it are where each chamber changes hands, and a tap on a candidate holds that race wherever the slider goes. The chances and election night are pretend runs that add a random miss for the whole country (about 3 points) and for each race (about 5), numbers I chose. The faded seats in the Senate are not on the ballot this year. Independents who sit with the Democrats count with them, and Republicans keep the Senate at 50 because the Vice President, JD Vance, breaks a 50-50 tie. This is a game built on the forecasters' numbers, not a forecast of mine.

The betting markets are what people with real money on the outcome think: Polymarket and Kalshi, read every hour. A price is a chance: 70 cents on a candidate pays a dollar if they win, so the market gives them 70 in 100. Where both price a race, the page shows the average. "Betting favorites" gives every race they price to the side they favor; a race they do not price keeps following the slider.