House Effects
Monday, September 21, 202643 days until Election Day, November 3, 2026
2026 projections

The Senate

Of the 35 seats on the ballot, Republicans hold 22. Democrats hold 34 seats that are not on the ballot and Republicans 31, so Democrats need to win 17 of the 35 for an outright majority, or 16 if an unaligned independent takes a seat from the Republicans: the party with more seats organizes the Senate, and the vice president only breaks ties. A national polling miss moves every race together in the simulations, which is why the seat distribution is wider than the race-by-race odds suggest.

Democrats win control
56%
Democrats 56%, Republicans 44%. The party with more seats organizes the Senate and the vice president only breaks ties, so Democrats need 51 seats, or 50 if an unaligned independent takes one from the Republicans (8% of simulations end that way). 9% are ties that the vice president settles for Republicans.
Democratic seats, expected
50.3
An average over simulations, not a single forecast. 47 now; nine in ten simulations: 46 to 54.
434547495153555751: outright Democratic majority →17%
Democrats control Republicans controlDemocratic seats after the election, 40,000 simulationsThe bar at 50 is split: Democrats control with 50 seats when an unaligned independent holds a seat, and Republicans control a 50-50 Senate through the vice president.
Safe D 8Likely D 5Lean D 3Toss-up 4Lean R 1Likely R 3Safe R 11
The assumption that matters most: which way the polls miss
If…Democrats win controlRepublicans win controlDemocratic seats
No direction assumed (still carries the usual uncertainty)56%44%50.3
Polls overstate Democrats as they did, on average, in 2016–24 (2.4 points)28%72%48.6
Polls understate Democrats by the same amount (2.4 points)81%19%51.9

“No direction assumed” does not mean the polls are assumed to be exactly right: every row above, including that one, already carries the historical size of past misses as uncertainty (see Uncertainty) — the win probabilities and seat ranges everywhere on this page are built from that spread. It means only that the model does not bet on which direction the next miss runs, unlike the two rows below it. From 2016 through 2024 the projections this method would have made ran too Democratic on average; before that they ran about even, and in the 2025 governor's races the polls missed the other way. These rows show what a repeat of that pattern, or its mirror image, would mean on top of the uncertainty already built in.

Race by race

All 35 seats on the ballot

Sorted from most Democratic to most Republican. The heavy rule is the line for control: Democrats need the race just above it, and every race above that. Where an independent stands in for a missing Democratic nominee, the margin is independent minus Republican, and a win there counts for neither party, which lowers the number Democrats need. See how the intervals are built.

Race
D+30EvenR+30
Proj.
projected margin (polls + fundamentals) polls alone half of simulated outcomes nine in ten outcomes
Table view

Every race, sortable

35 races
CandidatesRatingRaised (D / R)Cook
DelawareCoons v. KatzSAFE DD+27.0>99% DD+27.0D+16$1.5M / $12KSafe D
MassachusettsMarkey v. DeatonSAFE DD+26.2>99% DD+20.68D+29.6D+27$1.0M / $339KSafe D
Rhode IslandReed v. McKaySAFE DD+25.9>99% DD+19.53D+27.4D+15$1.6M / $55KSafe D
OregonMerkley v. SmithSAFE DD+24.699% DD+24.6D+16$2.1M / $75KSafe D
ColoradoHickenlooper v. BaisleySAFE DD+22.598% DD+22.5D+12$4.3M / $73KSafe D
VirginiaWarner v. MizusawaSAFE DD+20.8>99% DD+22.42D+20.1D+7$5.2M / $50KSafe D
New JerseyBooker v. MurphySAFE DD+20.797% DD+20.7D+7$5.0M / $7KSafe D
IllinoisStratton v. TracyLIKELY DD+16.994% DD+16.9D+12$3.7M / $368KSafe D
New MexicoLuján v. MarkerSAFE DD+16.499% DD+15.21D+16.6D+7$1.6M / $25KSafe D
GeorgiaOssoff v. CollinsLIKELY DD+8.894% DD+7.710D+10.6R+1$28.6M / $2.2MLean D
North CarolinaCooper v. WhatleyLIKELY DD+6.990% DD+8.630D+3.3R+2$6.6M / $4.0MLean D
New HampshirePappas v. SununuLIKELY DD+6.588% DD+4.922D+9.2D+4$5.6M / $1.4MTossup
MinnesotaFlanagan v. TafoyaLIKELY DD+5.383% DD+3.18D+7.9D+6$2.0M / $2.1MLikely D
MaineJackson v. CollinsLEAN DD+3.072% DD+2.311D+4.7D+8$6.2M / $6.3MTossup
MichiganEl-Sayed v. RogersLEAN DD+2.669% DD+2.231D+4.6Even$5.6M / $3.4MTossup
Ohio (sp.)Brown v. HustedLEAN DD+1.561% DD+3.620R+3.2R+10$17.6M / $5.1MTossup
TexasTalarico v. PaxtonTOSS-UPD+1.259% DD+2.736R+3.7R+12$29.5M / $11.3MTossup
NebraskaOsborn (I) v. RickettsTOSS-UPR+1.056% RR+1.06R+12.7R+19$2.3M / $2.1MLikely R
IowaTurek v. HinsonTOSS-UPR+1.360% RD+0.319R+5.6R+12$3.1M / $4.3MTossup
AlaskaPeltola v. SullivanTOSS-UPR+1.360% RD+2.312R+5.3R+12$7.7M / $2.6MTossup
KansasHamilton v. MarshallLEAN RR+4.680% RR+2.56R+6.8R+15$3.0M / $1.8MLikely R
Florida (sp.)Nixon v. MoodyLIKELY RR+7.491% RR+4.17R+11.0R+12$177K / $2.3MSafe R
South CarolinaAndrews v. GrahamLIKELY RR+7.691% RR+5.93R+8.9R+16$3.1M / $1.4MSafe R
South DakotaBengs (I) v. RoundsLIKELY RR+8.087% RR+8.06R+22.9R+28$315K / $1.1MSafe R
MississippiColom v. Hyde-SmithSAFE RR+12.197% RR+9.12R+13.1R+21$1.6M / $1.8MSafe R
LouisianaDavis v. LetlowSAFE RR+13.297% RR+7.51R+14.6R+21$151K / $977KSafe R
IdahoAchilles (I) v. RischSAFE RR+13.697% RR+13.66R+27.6R+35$461K / $1.9MSafe R
ArkansasShoffner v. CottonSAFE RR+15.6>99% RR+7.34R+21.7R+29$894K / $2.7MSafe R
MontanaBankhead v. AlmeSAFE RR+19.1>99% RR+21.511R+14.7R+18$27K / $730KSafe R
KentuckyBooker v. BarrSAFE RR+21.1>99% RR+12.72R+23.1R+29$192K / $4.3MSafe R
AlabamaWess v. MooreSAFE RR+23.2>99% RR+15.01R+25.2R+29$2K / $1.7MSafe R
OklahomaThomas v. HernSAFE RR+25.9>99% RR+24.01R+26.4R+33$14K / $1.0MSafe R
TennesseeBradshaw v. HagertySAFE RR+28.2>99% RR+24.61R+29.0R+28$0 / $2.7MSafe R
WyomingByrd v. HagemanSAFE RR+33.9>99% RR+33.9R+44$6K / $427KSafe R
West VirginiaAnderson v. CapitoSAFE RR+37.9>99% RR+37.9R+40$10K / $2.3MSafe R

Bold surname = incumbent running. Settled = only one major party is on the November ballot. Lean = presidential vote relative to the nation. Cook = Cook Political Report rating as listed on Wikipedia, shown for comparison only; it is not an input.