Midterms 2026, measured.
Kalshi's chance in every House, Senate and governor race and in both chamber-control markets, read every hour, and each race's status as it changes: 95%+ on the price after the polls close, called (Kalshi closed trading), settled, each with its time. These are the market's figures, dated; not a forecast of ours and not advice.
reading…
Control of Congress
reading…
Kalshi's control markets pay on the party of the Speaker of the House, and of the President pro tempore of the Senate, on February 1, 2027, and Kalshi may settle them earlier once at least four of the eight media outlets its terms name agree (Kalshi's contract terms, CONTROL).
The tally
reading…
"Called" here means one thing: Kalshi closed trading on the race. Kalshi's rules let it decide a winner market early once at least four of the eight media outlets they name (the AP, Decision Desk HQ, The New York Times, CNN, Fox News, NBC, CBS and ABC) have declared a winner and none disagrees (contract terms), and it closes trading when it does. Its close can come long after the polls close (New Jersey's 2025 governor race closed about 50 hours after its polls did), so the time shown is Kalshi's close, not the moment any outlet declared a winner; this site keeps no record of the outlets' own calls. "95%+ when polls closed" is the chance at the last hourly read in the two hours before the first polls in the state closed; "crossed 95%+ after" is the first hourly read after that at which an outcome reached 95% that was not already there. A race with no chance read in those two hours counts in neither. Before the first polls close (6:00 PM ET, Nov 3) both of those columns stay at zero.
Every race
reading…
Polls close: the hour the first polls close in the state, Eastern Time (in a state split across time zones the rest close up to an hour later). Kalshi's chance: each outcome's YES ask divided by the sum of the asks, to a tenth of a point. The 24-hour change is the leader's chance now less its chance at the last logged read at least a day old.
How it is measured
The races. Every 2026 House, Senate and governor race Kalshi lists a winner market for, from its own schedule of the races (its milestones, read once a day), and its two control markets. A race Kalshi lists without an entry in that schedule is added by name (on October 22, 2026, one: the Louisiana Senate race). A race whose market ticker names one state while its rules name another is held back until it is checked by hand, and counted above.
Kalshi's chance. Read every hour. Each outcome's YES ask, in dollars to the tenth of a cent, divided by the sum of the asks of every outcome still trading, to a tenth of a point. There is no chance that hour unless every outcome has an ask and the asks sum to between 97c and 125c. It is the market's own price, read, not a forecast.
The statuses. "95%+ when polls closed": at the last read in the two hours before the poll close, the leader was at 95% or more. "Crossed 95%+": the first read after the poll close at which an outcome was at 95% or more, unless it already was at the close; a race with no chance read in the two hours before the close has no reading at the close and no crossing is recorded for it; the reads are hourly, so the time is the read's ("by"). "Called": Kalshi closed trading on the race (every market of it stopped trading), at the close time Kalshi records, which can be hours or days after the polls close, and is not the time any outlet declared a winner. "Settled": Kalshi settled every market of the race, with its time and the outcome that settled Yes. The winner markets pay on the party (or the person) sworn in, and the control markets on the Speaker and the President pro tempore on February 1, 2027, but Kalshi may settle a market early once at least four of the eight outlets agree, as it did for New Jersey's governor race in 2025, two days after the election.
After November 3. The minutes from the poll close to 95%+, to Kalshi closing trading and to settlement, measured to the minute from Kalshi's own price history, and the races at 95%+ when the polls closed that settled the other way. The full method, fixed before the first polls close, is on the method page.
Where these numbers come from
Kalshi's public market data (its race schedule, its markets and their asks, close times and settlements), read by agents/midterms.py once an hour into the views of midterms.sql: midterm_races_public (each race shown), midterm_reads_public (the logged reads) and midterm_scan_public (the last read). Derived figures only: no ticker, bid, ask, size, volume or open interest is published, and nothing here is offered as a download.