How Long Til Trump Is Gone? · Midterm report card
The Trump Score
Provisional · includes 25 reported leads
A bruising result against the benchmarks
How Trump’s party is doing against pre-election expectations and modern midterm history. This editorial index uses probability curves; uncalled leaders count provisionally.
- Called results only
- 20/100
- Unresolved-race range
- 7–23
- HouseR 174 seatsExpected R 197 · net -4650% weight · called + provisional
- SenateR 44 seatsExpected R 47 · net -935% weight · called + provisional
- GovernorsR 20 seatsExpected R 22 · net -615% weight · called + provisional
481 sourced calls · 25 provisional leads · 0 races without a usable call or lead. Coverage: 506/506 races. 0 records outside the prior-party baseline excluded.
Source publication times used: Nov 8, 2026, 1:25 AM CST. 0 used records have an unavailable source time.
Dated shared snapshot built Nov 8, 2026, 1:26 AM CST. Live data replaces it when all three offices have usable coverage.
Benchmarks: Frozen Nov 3 forecast + 1962–2022 midterm history. 50 is middle-of-the-benchmarks performance.
How the Trump Score works
The score uses curved performance ranks, not fixed points per seat. Each office’s Republican seat total is ranked against Scrutinel’s frozen November 3 pre-election forecast. We approximate its published central estimate and 80% range with a logistic probability curve: F(x) = 1 / (1 + exp(−(x − center) / scale)). The lower and upper scales put the published range endpoints at the 10th and 90th percentiles. A forecast rank of 10 means roughly the bottom tenth of that approximate distribution, not a 10% probability of that exact seat count. For the House and Senate, that forecast rank is averaged equally with a historical rank fitted to the president’s party’s seat changes in 16 midterms, 1962–2022, from Brookings. The historical curve matches the sample mean and standard deviation. Governors use the forecast rank only. The three office ranks are averaged with House 50%, Senate 35%, and governors 15%, then rounded and bounded to 1–100. Those blend weights are editorial choices. Fifty is middle-of-the-benchmarks performance; holding every pre-election seat can score highly because midterm losses were expected.
| Office | Central estimate | Published 80% range | Current approximate ranks |
|---|---|---|---|
| House | R 197 | R 177–213 | 7/100 forecast · 13/100 history |
| Senate | R 47 | R 44–51 | 10/100 forecast · 5/100 history |
| Governors | R 22 | R 19–26 | 19/100 (forecast only) |
The published central estimate is treated as the median and the reported 80% range as equal-tailed, then fitted with a split logistic CDF. Those are modeling assumptions; this is an approximation, not the publisher’s raw simulation distribution. The governor forecast covers 36 races; eight Republican holdovers are added to all three published values.
President-party net seat changes, relative to the preceding general election, using Brookings’ immediate-election convention. Starting in 1962 keeps chamber sizes at 435 and 100. Current changes use the tracker’s pre-election party baseline; intervening vacancies or party changes can make the conventions differ. Governors have no historical component in this version.
Sourced calls take priority. In uncalled races, the actual leading candidate counts provisionally; candidates sharing a party are never combined. Ties, conflicting calls, missing races, and unusable counts keep the previous party only for this scoring calculation. The called-only score keeps every uncalled seat at its previous party. The possible range lets every uncalled or missing seat go Republican or non-Republican; it is a scenario range, not a probability interval. Independents count as non-Republican here, without inferring their caucus. This is an editorial index built from approximate marginal ranks, not a calibrated probability of Trump winning or a joint forecast. Historical elections are context rather than identical repetitions of 2026. Forecast uncertainty and historical election-to-election variation are kept separate. Benchmarks stay frozen; results and reported leaders update live.
The previous-party baseline and call attribution are linked in each race card. Source dates above describe the evidence used; build time does not make older data current. Scoring method version 2. These are research benchmarks from one forecast and one historical series, not a forecast consensus.
Trump Score 11 out of 100. A bruising result against the benchmarks. Provisional · includes 25 reported leads.
The 2026 Midterms · November 3
Every State.
Every Vote Counts.
Follow the races that shape the next two years.
Select a state for reported votes and the sources behind them.
2026 midterms · Called races
House: D 252, independent/other 0, undecided 22, R 161. 413 races called. Net party changes based on called races: D +42, R -42, I 0. 46 called flips. Baseline coverage 435/435 races; 0 calls with unknown baseline. Independents are separate; no caucus-control inference.
Source publication times: 2026-11-08T07:25:52.883Z. 0 races with publication time unavailable. Snapshot retrieved: 2026-11-08T07:26:05.404Z.
Senate: D 52, independent/other 2, undecided 2, R 44. 33 races called. Net party changes based on called races: D +7, R -7, I 0. 7 called flips. Baseline coverage 35/35 races; 0 calls with unknown baseline. Independents are separate; no caucus-control inference.
Source publication times: 2026-11-08T07:25:53.933Z. 0 races with publication time unavailable. Snapshot retrieved: 2026-11-08T07:26:04.713Z.
Governor: D 29, independent/other 0, undecided 1, R 20. 35 races called. Net party changes based on called races: D +5, R -5, I 0. 5 called flips. Baseline coverage 36/36 races; 0 calls with unknown baseline. Independents are separate; no caucus-control inference.
Source publication times: 2026-11-08T07:25:54.229Z. 0 races with publication time unavailable. Snapshot retrieved: 2026-11-08T07:26:05.238Z.
Static summary built 2026-11-08T07:26:05.457Z. Source dates above describe the data; build and retrieval times do not establish freshness. Includes seats not up this year. Calls are preliminary, not certification.
State races & official results · works without JavaScript
- Alabama races · Official results
- Alaska races · Official results
- Arizona races · Official results
- Arkansas races · Official results
- California races · Official results
- Colorado races · Official results
- Connecticut races · Official results
- Delaware races · Official results
- District of Columbia races · Official results
- Florida races · Official results
- Georgia races · Official results
- Hawaii races · Official results
- Idaho races · Official results
- Illinois races · Official results
- Indiana races · Official results
- Iowa races · Official results
- Kansas races · Official results
- Kentucky races · Official results
- Louisiana races · Official results
- Maine races · Official results
- Maryland races · Official results
- Massachusetts races · Official results
- Michigan races · Official results
- Minnesota races · Official results
- Mississippi races · Official results
- Missouri races · Official results
- Montana races · Official results
- Nebraska races · Official results
- Nevada races · Official results
- New Hampshire races · Official results
- New Jersey races · Official results
- New Mexico races · Official results
- New York races · Official results
- North Carolina races · Official results
- North Dakota races · Official results
- Ohio races · Official results
- Oklahoma races · Official results
- Oregon races · Official results
- Pennsylvania races · Official results
- Rhode Island races · Official results
- South Carolina races · Official results
- South Dakota races · Official results
- Tennessee races · Official results
- Texas races · Official results
- Utah races · Official results
- Vermont races · Official results
- Virginia races · Official results
- Washington races · Official results
- West Virginia races · Official results
- Wisconsin races · Official results
- Wyoming races · Official results
House Seats
218 for a majority · 435 seatsLoading sourced race calls…
Net party changes are based on sourced calls; leads do not count.
U.S. House
Race Calls & Reported VotesEach House district shows its called winner or reported vote margin. Select a district to jump to its race in the state panel.
Loading district boundaries. State totals are shown until the district map is ready.
Browse races as a list ↓Focus the map: +/− zoom, arrows pan, 0 resets, F fits the selected state. Drag with a mouse after zooming; on touch, use the pan buttons and scroll the page normally.
Loading state boundaries…
Uncalled margins & other map colors
House boundaries: U.S. Census Bureau · 2026 election districts. Select a district to jump to its race in the state panel.
—States with votes
—Races with votes
—Races in this feed
Results Sources & Connection Status
Agreement between feeds is a comparison of reports, not an independent audit. Shared upstreams are labeled. Unknown source times are never presented as fresh counts. The People’s Tally · CC BY 4.0 and Elections Archive · licence & original-source exceptions. Original publishers are credited in each race.
Nationwide coverage from civicAPI, state counts through Open America, plus the available sources listed above. Open a race’s source comparison to see agreement, differences and original publishers. Counts remain preliminary.
Browse races nationwide
The map’s list alternative. Select an office above, then search state, candidate or district. Reported margins compare the actual leading candidate and runner-up, including independents; they describe reported votes, not forecasts.
Checking race sources…
Since you last marked as seen
Comparisons stay in this browser for up to 30 days. Mark as seen to save a baseline. Refreshing and automatic checks keep unread changes.
New calls, removed calls, changed leaders and count corrections are separate observations. A change of count source is shown without a vote-change alert. This history does not determine results.
Read the Results, Keep the Context
A Count Is a Snapshot.
It Can Still Change.
What Do the Colors Mean?
The House map colors each congressional district separately: red for a called Republican winner, blue for a called Democratic winner, and teal for another party. Select a district to focus its race in the state panel. Senate and governor maps use state totals; states with called winners from different parties have stripes. A call takes precedence over a close or different reported vote lead.
In every individual race, uncalled margins compare the actual leading candidate with the runner-up, divided by all reported votes. Candidates from the same party are never combined into one opponent. Purple means the leading candidate is within three percentage points of the runner-up. House map colors use each district’s own counts. Separate regular and special Senate races are combined for the state map. House state-panel summaries still combine the state’s district counts. In an individual race, teal means an actual independent or third-party candidate leads. For states with multiple listed races, the map aggregates reported party votes. These margins describe counted votes; we do not use them to call races.
Gold state badges appear on the Senate and governor maps: S for Senate and G for governor. If district boundaries cannot load, the House view falls back to state totals with H badges. A check means all listed races are called; a fraction means some are called. A separate small clock means call-source updates are delayed. A delay does not retract a call, and mixed winners do not imply a statewide party win.
Why Are Some States Gray?
The national feed lists all 435 voting House districts, 35 Senate contests and 36 state governor races for November 3. Not every state elects a senator or governor this year. States marked “Not up for election” show their sitting senators or governor when selected, including party, next regular election year and source links. Senate states with one race also list the other senator whose seat is not up. Those seats are already included in the Senate counter and never counted as new calls. Light gray means a listed race has no counts; hatched gray means no race results are available in the current feeds. Neither means a tied race. Use the official state results link for the full ballot, including D.C.'s delegate, state legislative races, local contests and ballot measures.
Is Reporting Percentage the Percentage of Votes Counted?
No. States report different units, such as precincts, counties or localities. Some supply a statewide figure even for district races. We show the source's reporting figure on each race, without calling it the percentage of ballots counted or averaging it across races. A figure of 100% does not establish certification or a winner.
What Does the ⇄ Flipped-Seat Marker Mean?
A flip means a sourced call names a winner from a different party than the party that held that seat before this election. The orange ⇄ state badge counts flips on the Senate and governor maps; the House map shows district colors instead. Race cards in every tab show the party change, such as R → D. A lead alone never triggers a flip. A retirement followed by a same-party winner is a hold. Conflicting calls remain unresolved, and retractions remove the marker.
The prior-party baseline comes from the Elections Archive 2026 election catalogue (CC BY 4.0), retrieved November 4, 2026. Occupied House seats were cross-checked against the official Clerk roster. Each flip links to its prior-party source and its winning call. A flip is a reported party change, not official certification.
When Is a Race Called or Certified?
We mark a race called when civicAPI explicitly identifies one winner, or a linked, named newsroom call is supplied by The People’s Tally. Every call identifies its caller and source link. Conflicting calls remain unresolved and are excluded from the seat counters. civicAPI’s decision desk uses its published calling criteria. We accept that third-party call without applying our own reporting-percentage or margin formula. Only a later provider retraction removes the call; delayed updates remain labeled separately. Visit AP or Decision Desk HQ for their own calls. Calls and official certification are distinct; the EAC explains certification.
How Fresh Are the Numbers?
We check results about once a minute while this page is visible. The House and Senate seat counters update with those checks and use explicit calls, never vote leads. Senate totals include the 65 seats not up this year (32 Democratic, 31 Republican and 2 independent), based on the official Senate class lists; House totals cover all 435 voting seats. Source comparisons show each report’s publication time when available. Retrieval time is separate; an unknown publication time does not establish freshness. National feed timestamps are not per-state update times. Minnesota publishes every 15 minutes or less, and other sources have their own cadence. Your browser restores previously retrieved results immediately while checking for updates. A shared backup cache is scheduled to refresh every five minutes during election week, though schedules and source delivery can be delayed. Older responses cannot replace newer cached data. Among complete reports, we prefer the newest known publication time, favoring direct official sources when times match. civicAPI supplies the national race catalogue and explicit calls; additional named calls remain attributed to their newsroom. We select one whole record per race without combining providers’ vote counts. Cached or delayed records retain their source dates and are labeled. Official state sources remain available to verify the figures.