Democrats are clear favourites to take the House and modest favourites to take the Senate. The House follows from a lopsided national environment. The Senate comes down to seven close races, mostly in states Trump won comfortably in 2024. All probabilities marked ours blend markets, models, polls and ratings, then run through a simulation.
Senate
Now 53 R – 47 D- Most likely result
- 52 D – 48 R
- Democratic seats, 80% range
- 47 – 55
- Prediction markets
- ~65% D
- Forecast models
- 55 – 75% D
House
Now 218 R – 214 D – 1 I · 2 vacant- Central estimate
- 231 D seats
- 80% range
- 216 – 246
- Prediction markets
- ~91% D
- Forecast models
- 77 – 94% D
Projected Senate
Each race given to its favourite · tap a seat group below for detailThe seven races that decide it
Share of simulations where this race delivers the 51st seatNo single race is decisive; each of these is the tipping point about 10% of the time. North Carolina sits underneath them all: if Democrats lose it, their Senate chance falls to about 8%.
The environment
Latest readings, October 2026Polls understated Republicans by roughly 3–4 points in 2016, 2020 and 2024. A 3-point miss in their favour cuts Democratic Senate odds to 29% and House odds to 62%. A 3-point miss the other way, in line with 2025–26 special elections, lifts the Senate to 84%.
Markets price Democratic control near 65%. We land at 59% because the consensus ratings still call most decisive races toss-ups, and because a 1–3 point lead in Texas, Ohio, Alaska or Maine is well within normal polling error. We agree with the markets on the most likely map.
The math
- The Senate is 53 R – 47 D (45 Democrats plus 2 independents who caucus with them).
- 35 seats are up: 33 regular plus specials in Ohio and Florida. Republicans defend 22, Democrats 13.
- Seats not up: 31 R, 34 D/I. Democrats need 17 of the 35 to reach 51. A 50–50 Senate stays Republican because VP Vance breaks ties.
- If every race goes to its favourite, Democrats flip NC, AK, TX, OH and ME and hold all 13 of their seats: 52–48.
Democratic seat total
Share of 50,000 simulated elections ours51+ seats gives Democrats control. At exactly 50, Republicans keep control through the vice president, unless Osborn wins Nebraska and sides with Democrats.
The 17 contested races
Tap a race for polls, money, ratings and our reasoningWhat each race is worth
Democratic control chance if they lose or win the race ours
The Nebraska wildcard
- No Democrat is on the ballot. Nebraska Democrats back independent Dan Osborn against Sen. Pete Ricketts.
- Polls range from Osborn +4 to Ricketts +5; average Ricketts 46.0 – 43.7.
- Osborn wins in about 30% of our simulations. In about 2%, the chamber lands at 50 D – 49 R – Osborn, and his choice decides control.
- We found no statement from Osborn on which party he would caucus with.
- Money is roughly even: $12.8M from pro-Osborn groups through Oct 13; $15M from the GOP's Senate Leadership Fund.
Colorado (Hickenlooper), Delaware (Coons), Illinois (Stratton, open), Massachusetts (Markey), New Jersey (Booker), New Mexico (Luján), Oregon (Merkley), Rhode Island (Reed), Virginia (Warner)
Alabama (Moore, open), Arkansas (Cotton), Idaho (Risch), Louisiana (Letlow, beat Sen. Cassidy in primary), Oklahoma (Hern, open), South Dakota (Rounds), Tennessee (Hagerty), West Virginia (Capito), Wyoming (Hageman, open)
How we built each race probability
Each race blends four inputs. Weights are our own choice and are rescaled when an input is missing.
- Prediction markets, 35%. Kalshi on Oct 21, averaged with Polymarket where it trades.
- Models, 25%. Silver Bulletin race odds (about Oct 6, the latest public) and Decision Desk HQ (Oct 21).
- Polling average, 25%. Converted to a probability assuming a 5.5-point normal error.
- Ratings, 15%. Cook, Inside Elections and Sabato, mapped Toss-up 50%, Tilt 55%, Lean 65%, Likely 85%, Solid 99%.
The simulation adds a shared national swing (SD 3.5 pts) and independent state noise (SD 4.0 pts), so races move together at a correlation of about 0.43. No directional "polls underrate Republicans" adjustment is applied; both directions are shown as scenarios instead.
House control
218 needed- Central estimate
- 231 D seats
- 80% range
- 216 – 246
- If polls miss 3 pts toward GOP
- 221 seats · 62%
- If polls miss 3 pts toward Dems
- 241 seats · 97.6%
Starting point
- Current House: 218 R, 214 D, 1 independent (Kevin Kiley, ex-Republican), 2 vacancies (FL-20, TX-23). Other counts: 219–214, or 220–215 at some outlets. The target is 218 either way.
- Democrats need a net gain of about 4 seats.
- 61 members are leaving, the most since 1992. Republicans set a single-party retirement record.
- Since 1946 the president's party has lost about 25 House seats on average, and about 37 when his approval is under 50%.
Cook Political Report ratings
Oct 16 · 435 seats214 seats lean Democratic or better, so Democrats need about 4 of the 21 toss-ups. 16 of those toss-ups are Republican-held. In October 2018 Cook had 45 GOP seats at toss-up or worse; today it has 23, so this board is wave-like but with fewer exposed GOP seats.
Mid-decade redistricting
| State | Expected effect | Status |
|---|---|---|
| Texas | R +3 to +5 | In effect (SCOTUS allowed, Dec 2025) |
| Florida | R +1 to +4 | In effect; merits case pending |
| Ohio | R up to +2 | In effect |
| North Carolina | R +1 | In effect |
| Tennessee | R +1 | In effect (Memphis seat split) |
| Alabama | R +1 | In effect (injunction lifted Jun 2) |
| Louisiana | R +1 | In effect after Callais; Dec 12 runoff |
| California | D +3 to +5 | In effect (Prop 50) |
| Utah | D +1 | Court-ordered map |
| Missouri | None | Blocked by SCOTUS Sep 25; Prop A on ballot |
| Virginia | None | Dem amendment voided May 8 |
- On paper (maps in effect)
- ~ R +9
- Cook estimate
- ~ R +4
- GOP strategists (Axios, Oct 20)
- R +1 to +6
- Our estimate of realized net
- R +3 to +5
Several new GOP-drawn seats are themselves competitive (FL-14, FL-22, OH-09, NC-01, TX-34). Strategists expect only 2–4 of the 5 Texas targets and have pulled out of TX-28. A D+8 generic ballot covers this cushion several times over.
Key districts
Cook toss-ups plus the closest model seats · lean is ours| District | Held by | Signals | Our lean |
|---|
Late market movers
GOP incumbents outside Cook's toss-up tier · 72-hour change| District | Incumbent | GOP win price | 72h change |
|---|---|---|---|
| OH-10 | Turner | 65% | −18.3 |
| KY-6 | Alvarado | 56% | −11.8 |
| NJ-2 | Van Drew | 72% | −11.5 |
| FL-4 | Bean | 78% | −8.5 |
Shift the national polling error and fix the result of any race. The page reruns 50,000 simulated elections with the same model behind our forecast, keeping only the runs that match what you fixed.
Fix a race
D / model / RMontana (three-way race, 96% Alme) and the 18 safe seats stay at the model.
The fundamentals look like 2006 or 2018, both years when the president's party lost the House badly. The president is near his second-term low, the generic ballot has widened all year, and voters rate the economy harshly.
Generic ballot
Democratic lead, pointsOctober polls: NYT/Siena D+11, CNN D+10, NBC D+10, Reuters D+12, Fox D+9, AP-NORC D+9, Rasmussen D+4. Democrats won the House vote by 8.4 in 2018 and gained 41 seats.
Trump net approval
Approve minus disapproveLatest: RCP 38.2 / 60.5. Pew 30%, Reuters/Ipsos 31% (record low), CNN 32%, NYT/Siena 36%, Fox 36%. Approval among his 2024 voters is down to 67% (Pew).
Economy
Latest official readingsEvents that shaped the year
- Iran war. US and Israel struck Iran on Feb 28; Iran closed the Strait of Hormuz. A ceasefire collapsed Jul 8. Fighting was ongoing in late September; October status unconfirmed.
- Shutdown and health care. Record 43-day shutdown in Oct–Nov 2025. Enhanced ACA subsidies expired Jan 1, 2026.
- Tariffs. Supreme Court ruled 6–3 on Feb 20 that emergency powers do not authorize them, opening up to ~$175B in refunds.
- Immigration enforcement is drawing pushback; about half the Epstein files have been released.
- Government is funded through Dec 11, setting up a lame-duck fight.
Issue standing
| Measure | Reading |
|---|---|
| Trump net on inflation (Silver) | −48.9 |
| Trump net on economy (Silver) | −32.9 |
| Trump net on trade (Silver) | −30.6 |
| Cost of living approval (ABC/WaPo) | 21 / 77 |
| Trust on health care, D v R (NYT/Siena) | 60 – 36 |
Republicans campaign on no tax on tips and overtime and on drug pricing. Democrats run on costs and health care.
Behaviour signals
- Special elections: Democrats beat the 2024 margin by about 13 pts on average since 2025 (about 11 in 2026 specials). 10 red-to-blue flips, none the other way.
- 2025 governors: Sherrill won New Jersey by 14.4, Spanberger won Virginia by 15.4.
- Enthusiasm: 48% of Harris voters "very enthusiastic" against 30% of Trump voters (Reuters/Ipsos).
- Early vote: ~3.6M ballots by Oct 13; Michigan absentee requests up 22% on 2022.
Specials overstate swings because turnout is low, so we treat these as support for the polls, not a reason to add to them.
Money, Q3
| Who | Dem side | GOP side |
|---|---|---|
| Senate committees (DSCC v NRSC) | $35M | ~$25M |
| House committees, cycle (DCCC v NRCC) | $249M | $250M |
| Senate super PACs (SMP v SLF) | $202M | $273M |
| House super PACs (HMP v CLF+) | $230M | ~$94M |
| Texas Senate candidates | $64M | $10.2M |
Elon Musk gave $178M to the two main GOP super PACs in Q3. Democratic candidates outraised Republicans in most competitive Senate races; GOP outside money narrows that gap.
Every forecast favours Democrats in the House. On the Senate they split by 20 points, mostly by method: poll-driven models read small leads in Texas, Ohio and Alaska as real; fundamentals-heavy models discount them. Markets sit in between.
Democratic chance by forecaster
Senate ● and House ◆ · ours highlighted| Forecast | House D | House seats | Senate D | Senate seats | As of |
|---|---|---|---|---|---|
| Silver Bulletin | 93.6% | 238.7 avg | 75% | 52.2 avg | Oct 19 |
| Split Ticket | 89% | – | 65% | – | Oct 14 |
| Poliwave | 88% | 231 (217–249) | 55% | 51 (48–54) | Oct 21 |
| Call the Map | 82% | – | 56% | – | Oct 18 |
| Decision Desk HQ | 77% | 235 | 55% | 51 | Oct 21 |
| The Economist | – | – | 60% | 51 | Oct 7 |
| FiftyPlusOne | 99% | – | 73% | 52 median | Oct 1 |
| Race to the WH | – | – | 68.9% | – | Sep 22 |
| Kalshi | 90.8% | – | 65% | – | Oct 21 |
| Polymarket | 92% | – | 64.5–65% | – | Oct 21 |
| Ours | 87.2% | 231 (216–246) | 59% | 51 median (47–55) | Oct 21 |
Senate odds over time
Democratic chance of controlPolymarket's August point is an approximate mid-month placement. Drivers cited: strong Democratic Senate polls in Texas, Ohio and Michigan from mid-September, then Maine, Alaska and Kansas; rating shifts by Sabato (Sep 22) and Cook (Oct 6); generic ballot near D+9.
Both chambers
| Outcome | Ours | Polymarket | DDHQ |
|---|---|---|---|
| Democratic sweep | 59.0% | 62.5–64% | 54% |
| D House, R Senate | 28.2% | 27.5–28% | 24% |
| Republican sweep | 12.6% | 8.5–9% | 21% |
| D Senate, R House | 0.2% | <1% | – |
Where they disagree
- Alaska: markets 74–76% Peltola; models 61–65%.
- Texas: Silver 74% Talarico, markets 64%, DDHQ 55%.
- Iowa: polls tied and DDHQ has Turek 55%, but markets give Hinson 56–61%.
- Kansas: Silver called it a pure toss-up; DDHQ gives Marshall 65%; markets ~60%.
- Maine: Kalshi 64% Jackson, Polymarket 56%.
- DDHQ's model partly uses market data, so the two are not fully independent. The Senate market is thin: $448K traded on Kalshi in 24 hours.
How Republicans keep the Senate
~39% ours- Hold North Carolina's open seat (Democrats are 92% there; losing it leaves them ~8% for control).
- Or win any two of Maine, Ohio, Texas, Alaska and Michigan while keeping Iowa and Kansas.
- Candidates who break with Trump: Rogers on Canada tariffs and Iran, Sununu on Iran, Collins's long record of independence.
- Outside money: pro-GOP groups lead ad spending about 58/42 in Michigan and Maine; over $200M of national GOP money, nearly all from Musk, has gone into Texas.
- A repeat of the 2016–2024 polling miss.
Late-breaking factors
- Iran war. A ceasefire or a reopened Strait of Hormuz could ease gas prices and help Republicans at the margin; escalation would do the opposite.
- Gas prices. About $4.27 now against a $4.56 record in May. Partial relief is the clearest route to a GOP recovery.
- Turnout. Early signals lean Democratic: Democrats are 67% of Iowa mail-ballot requests (54% in 2024); Ohio Republicans are reported worried about turnout.
- Trump's travel. He is campaigning for Marshall, Ricketts and Sullivan. In Kansas his net approval is about even in one poll.
Election night and after
Control may not be known on Nov 3Data caveats
- Silver Bulletin race-level odds are from about Oct 6; newer figures are paywalled.
- The status of the Iran war in October could not be confirmed.
- Decision Desk HQ pages showed what looked like placeholder pollster entries; only their averages were used. DDHQ's "70% R" South Carolina reading was excluded.
- House composition differs by source (218–214, 219–214, 220–215).
- Kalshi showed Maine at 57% on Oct 16 and 64% on Oct 21; we could not confirm whether that was a rebound or a quoting difference.
- Model figures mostly come via news coverage because model sites are paywalled or chart-only.