← Back to the election briefing
Forecast updates
What changed in the model, and when. Explore each published version alongside its recorded revisions.
These are revisions to published estimates, not a measure of changing voter opinion. New polls, corrections and model inputs can change a forecast; a recorded reason is shown when one is available. All times are Central.
The forecast, version by version
18 published versions from Nov 1, 2:50 am CT to Nov 3, 12:24 pm CT. Each point records a published model version.
House — solidSenate — dashed
Vertical position: chance Democrats end the Republican supermajority. Higher means that outcome is more likely.
Nov 3, 12:24 pm CT
House: 89%Senate: 54%
Chance Democrats end the Republican supermajority
Recorded changes
No top-line chance moved by a whole percentage point; district-level figures may have changed.
These are changes to the model’s published estimates. They do not by themselves show a change in voter opinion. No cause was recorded for this revision.
All published values and sources
| Published (Central) | House | Senate |
|---|---|---|
| Nov 1, 2:50 am CT | 87% | 43% |
| Nov 1, 10:53 am CT | 90% | 48% |
| Nov 1, 11:40 am CT | 90% | 48% |
| Nov 1, 1:33 pm CT | 90% | 48% |
| Nov 1, 2:40 pm CT | 90% | 48% |
| Nov 1, 3:08 pm CT | 90% | 48% |
| Nov 1, 3:56 pm CT | 90% | 48% |
| Nov 1, 7:29 pm CT | 90% | 48% |
| Nov 2, 3:22 am CT | 90% | 48% |
| Nov 2, 11:28 am CT | 91% | 54% |
| Nov 2, 11:57 am CT | 91% | 54% |
| Nov 2, 12:04 pm CT | 91% | 54% |
| Nov 2, 2:18 pm CT | 91% | 54% |
| Nov 2, 8:53 pm CT | 91% | 54% |
| Nov 2, 10:10 pm CT | 91% | 54% |
| Nov 2, 10:35 pm CT | 89% | 54% |
| Nov 3, 4:44 am CT | 89% | 54% |
| Nov 3, 12:24 pm CT | 89% | 54% |
Source: every published version of this site’s forecast. Method · Download history data.
Decisions behind the changes
Amendment 3: the prior no longer counts the polls twice
Amendment 3's starting point is now the 2024 vote carried to the 2026 electorate and the statewide shift, with no poll-based bump, and it is allowed a little more uncertainty. The polls enter once, as polls.
Published in the version of Nov 2, 10:35 pm CT:
- Amendment 3: Passes 35% → 30% (down 5 points)
Wave-year uncertainty: seats are allowed to move together more
Regional and district-type shared errors are set to the 2018 wave level, so seats can swing together more. No point estimate moves; Democrats' chance to end the House supermajority reads about 89% instead of 91%.
Published in the version of Nov 2, 10:35 pm CT:
- Missouri House: Republican supermajority ends 91% → 89% (down 2 points)
Wave-year uncertainty: seats are allowed to move together more
Regional and district-type shared errors are set to the 2018 wave level, so seats can swing together more. No point estimate moves; Democrats' chance to end the House supermajority reads about 89% instead of 91%.
Not yet in a published version; the hourly run publishes the new figures.
Amendment 3: undecided voters now split 22% to Yes
Amendment 3's undecided voters are now split 22% to Yes instead of 39%. In 12 of the 14 abortion measures of 2022 to 2024 with final-month polls, the status quo did better than its poll. The site's recommendation is editorial and unchanged.
Published in the version of Nov 2, 10:10 pm CT:
- Amendment 3: Passes 40% → 35% (down 5 points)
Petition measures: undecided voters now split 22% to Yes
On measures put on the ballot by petition (Amendment 6, Proposition A), undecided voters are now split 22% to Yes instead of 39%. Across 68 statewide measures since 2014, the result fell short of the final two-way poll in 79% of contests and in 86% of citizen initiatives.
Published in the version of Nov 2, 10:10 pm CT:
- Amendment 6: Passes 79% → 70% (down 9 points)
- Proposition A: Passes 29% → 20% (down 9 points)
Election night: early absentee votes counted as their own batch
The election-night model now treats each county's early absentee batch as its own votes (in 2024 these ran a median 4.3 points more Democratic than election-day votes) and marks a seat Called only when its count is complete or at 99% or more with at least 15% of its vote counted, a rule scored on a replay of the 2024 House count. No pre-election figure changed.
Election-night model; the pre-election forecast is not affected.