Senate Race Report: October 25, 2026

Question: Nine days before Election Day, Republicans are spending $26 million to defend Kansas, suing to block certification in Alaska and retreating from North Carolina, so why is control of the Senate still a coin flip?

Answer: Because every one of those facts describes a party playing defense, and defense is not the same thing as defeat. Cook Political Report moved three Senate ratings on October 20: South Carolina from likely to lean Republican, Mississippi from solid to likely Republican, and Minnesota from likely to lean Democratic. Silver Bulletin moved Iowa from toss-up to lean Democratic a day later.¹ That is four moves in seven days, three toward the Democrats, and together they describe a map that has widened rather than a majority that has settled. Kalshi put Democratic control at 68 percent on October 23, up from 62 percent on October 15 and from less than 43 percent on June 24.² The market is not hesitating about direction. It is hesitating about the last mile, which runs through five states that the forecasters describe in five different ways, and nine days is enough time for the last mile to move either way.

What Changed This Week

Kansas became the most expensive Senate race in its history, and the most revealing. AdImpact data reported by the Associated Press puts Republican group commitments at more than $26 million since the August primary, more than they spent in Marshall’s entire 2020 race. With Marshall’s campaign included, Republican spending passes $32 million, while Hamilton and Democratic groups have committed more than $38 million. More than 40 outside groups have spent about $71 million in a state that has not elected a Democratic senator since 1932, a Kansas record.³ The fundraising is stranger still. Since the primary, Hamilton raised $8 million nationally, including more than 10,000 contributions from Kansans, while Marshall raised $905,000, with about 200 of those contributions from Kansas residents.³ An incumbent senator raised roughly one ninth of what his challenger raised, in his own state, in the final ten weeks.

The last debate was a contest over which man could stand closer to, or further from, the president. Marshall said, “Control of the Senate goes through Kansas,” and tied Hamilton to Chuck Schumer, whose aligned super PAC has put $5 million into the race. Hamilton answered that he is not running against the president, and that Marshall “says he speaks up for us, but the president doesn’t seem to listen to what he has to say.”⁴ The sharpest exchange came over Marshall’s ad calling Hamilton a “pedophile protector.” Hamilton conceded a 2006 lapse, saying, “That was a mistake on my part,” and added, “Ultimately, we helped get (the perpetrator) put behind bars.”⁴ Two days earlier the president rallied in Park City and called Hamilton “the most radical liberal Democrat to ever get through a Senate primary in the history of Kansas.”³

Alaska stopped being a polling story and became a legal one. The Justice Department sued the state on October 22, asking a judge to allow a delay in certifying the election until an “investigation” determines whether ballots cast for Daniel J. Sullivan Jr. should have counted for the incumbent, and to block certification if the senator loses by fewer votes than the second Sullivan received. The complaint argues there is a “substantial likelihood” that certification “will reflect an electoral process that lacked integrity.”⁵ The president told reporters he ordered it himself: “I’m the one that said do it.”⁵ He had posted a day earlier that his administration was “looking at powerful legal options to remove the fraudulent Daniel J. Sullivan Jr. from the Ballot.”⁶ The market did not move. Kalshi still priced Peltola at 75 percent and Sullivan at 27 percent the following day.⁷ A lawsuit filed nine days before an election is less a legal argument than a message to the voters still deciding, and the message is that one side is preparing not to accept the count.

The raters moved in three directions at once. Cook cut South Carolina to lean Republican and Mississippi to likely Republican, and marked Minnesota down to lean Democratic. Silver Bulletin moved Iowa to lean Democratic.¹ Mississippi is the one that would have been unthinkable a month ago: Cook had it solid Republican, the president carried it by 23 points, and no Democrat has won a Senate seat there since 1982.⁸ In the same week, the Senate Majority PAC put $1.2 million into television and digital advertising there through Election Day, in the reddest state the group has contested.⁸

North Carolina stopped being a contest. NBC News reported on October 22 that Republicans are retreating from the state, with the president seeming to concede that Roy Cooper will win the seat and downballot Republicans “extremely nervous” about what that means for them.⁹ The Carolina Journal poll released the same day put Cooper ahead 49.4 to 39.1, a 10.3-point lead narrowed from 15.1 points in September, with Whatley gaining five points while Cooper stayed flat and undecideds falling from 12.3 to 6.8 percent.¹⁰ A narrowing lead in a race the other party has abandoned is not a comeback. It is the sound of the remaining undecideds choosing between a campaign that is still talking to them and one that is not.

The national numbers got worse for the party in power. The Strength in Numbers survey published October 21 put Democrats ahead 52 to 41 among registered voters and 54 to 42 among likely voters, the largest lead in the 17 waves of that survey since May 2025, with the president at 34 percent approval and 62 percent disapproval.¹¹ Early voting is running well ahead of the last midterm. An Associated Press analysis published October 24 counted more than 13 million ballots cast, against 9 million at the same point in 2022, with Ohio up 156,000 votes, a 30 percent increase.¹² Early votes are not results, and Republicans have historically voted later. But a 30 percent increase in a state with a competitive Senate race decides whether the final week is about persuasion or about turnout.

The Defense

Michigan remains the most competitive Democratic-held seat and the one where the closing argument has turned personal. In the second and final debate on October 21, El-Sayed called Rogers “a sock puppet for the man,” and Rogers answered, “He wants to make this about Donald Trump. The bingo game ought to be: How many times can he put Trump in any answer?”¹³ Rogers accused El-Sayed of a “pattern of deception,” and El-Sayed replied, “It doesn’t make any sense, and it’s actually really quite beneath you, Mike.”¹³ The polling has not moved much. InsiderAdvantage had El-Sayed ahead 47 to 46 in a survey completed October 22, and Kalshi priced him at 70 percent the same week.¹⁴¹⁵ A one-point race with a two-point error bar is a coin flip, and the market’s 70 percent is a judgment about the state’s Democratic lean as much as about the candidate.

Georgia is the defense seat that has stopped requiring defense. The Atlanta Journal-Constitution and University of Georgia poll, in the field October 7 to 15 with 1,000 likely voters, put Ossoff ahead 48 to 40 with nearly 12 percent undecided.¹⁶ Ossoff spent the week pressing Collins over his “Convoy Captains” volunteer program, saying, “Mike Collins needs to disband his so-called convoy captains’ program today,” and alleging that some of its leaders “have ties to the Ku Klux Klan and have criminal histories.”¹⁷ Kalshi priced Ossoff at 95.6 percent, with a polling average of 8.3 points in his favor.¹⁰ The remaining question is not whether he wins but whether the margin carries the rest of the ticket, and an eight-point lead usually does.

Minnesota is the Democratic seat moving the wrong way. Cook cut it from likely to lean Democratic on October 20, following Sabato’s move in the same direction six days earlier, which makes Minnesota the one Democratic seat that two forecasters have marked down while the national environment was improving for the party.¹ Rasmussen, in the field October 20 and 21 with 1,182 likely voters, had Flanagan ahead 48 to 45 with 7 percent unsure, unchanged from two weeks earlier, and Tafoya ahead among unaffiliated voters 46 to 43. The gender gap was 23 points, with women for Flanagan 54 to 40 and men for Tafoya 51 to 42.¹⁸ Flanagan is still the favorite. But a race that was expected to be comfortable now demands a final week, and every dollar spent in Minnesota is a dollar not spent in Ohio.

New Hampshire’s closing messages have already hardened. Sununu told WCAX, “It’s about the future. You know, not next year, but the next six years,” while Pappas said the race is “a way that we are going to be able to render a verdict on his reckless and incompetent leadership.”¹⁹ The polling has been stable and modestly Democratic all month. The last market price Kalshi published, on October 9, was 90 percent for Pappas.²⁰

The other nine Democratic seats, Colorado, Delaware, Illinois, Massachusetts, New Jersey, New Mexico, Oregon, Rhode Island and Virginia, are not in play.

The Offense, Tiered

North Carolina remains the anchor of every path, and this week it became the anchor of the Republican retreat as well. Cooper led 49.4 to 39.1 in the Carolina Journal survey.¹⁰ An Elon University poll had him ahead by 13, and Kalshi’s own average is 9.8 points, which is what happens when a state stops being contested and the pollsters stop pretending otherwise.¹⁰ Kalshi priced Cooper at 96.1 percent.¹⁰ The Republican nominee is still campaigning, but a party that believed it could hold this seat would not have conceded it aloud with nine days to go.

Ohio is the clearest case of a race where the national environment and the candidate pull in the same direction. Fox News, in the field October 15 to 19 with 1,592 likely voters, had Brown ahead 52 to 47, with Husted 13 points underwater on favorability, 42 percent favorable to 55 percent unfavorable, and an 18-point gender gap.²¹ The early vote is the other half of the story. Eight days into voting, Democrats had cast 41 percent of the nearly 230,000 ballots and Republicans 28 percent, against 26 and 36 at the same stage in 2024.²² Kalshi had Brown at about 60 percent on October 15.²² A strategist’s line in that report is the honest summary of the other side’s problem: “Anytime you win everything, you got a turnout problem.”²²

Texas is the race where the Democratic lead is real and shrinking. Texas Public Opinion Research, in the field October 15 to 18 with 969 likely voters, had Talarico ahead 51 to 47, a four-point lead that has narrowed one point a month since August, from six to five to four.²³ Talarico led independents 66 to 28 and women 54 to 47, while Paxton led men and white voters; the coalition that produced a four-point lead is not the coalition that usually elects a Texas Democrat, which is either the story of the year or the reason the lead is soft.²³ Kalshi priced Talarico at 64 percent on October 21.²⁴ There has still been no debate. The Texas Tribune reported this month that the race could be the first in Texas without one in a quarter century, with Talarico accepting at least four invitations and Paxton, who has avoided debates through three attorney general campaigns, accepting none.²⁵ On the first day of early voting the two held dueling rallies in the Fort Worth area, Paxton with Robert F. Kennedy Jr. and Talarico with the cast of “Friday Night Lights.”²⁶

Maine is now a tied race with a temperature problem. The Fox News poll, in the field October 16 to 20 and the first conducted entirely after the final debate, put Jackson and Collins at 49 apiece, with Jackson ahead by 40 points among the one in five who had already voted and Collins ahead by 13 among independents.²⁷ Joe Manchin campaigned with Collins in Lewiston and Norway on October 21, calling her “the soul of the Senate” and saying, “I have been through three presidents with her.”²⁸ Collins used the visit to emphasize her break with the administration on Canada, saying, “I have led the opposition to the imposition of Canadian tariffs.” The administration began pulling a batch of Canadian goods out of its 50 percent tariffs that week, and Collins’s office said an order would exempt nine categories of products Maine farmers rely on, though no proclamation had been issued.²⁸²⁹ Heating oil in Maine is up roughly 80 percent over last year, which is the kind of number that a five-presidents argument does not answer.²⁸ Kalshi had Jackson at 65 percent on October 23.²⁹

Alaska has a lawsuit, a ranked-choice count and a polling lead, in that order of certainty. Peltola has led most public surveys since the summer, and the last published market price put her at 75 percent; the DOJ suit is an attempt to make the count itself the issue.⁷ Under ranked choice, the second Dan Sullivan’s primary vote total matters only if the final margin is smaller than it, which is precisely the condition the complaint asks a judge to police.⁵ Alaskans are already voting. A federal court cannot un-cast those ballots, and the complaint asks it to do no such thing.

Iowa is the race where the models and the market have stopped agreeing. Kalshi had Hinson at 56 percent on October 20, after a final debate in which she said, “We need to lower gas prices in this country,” and he answered, “We are in this situation because of this unnecessary war in Iran.”³⁰ InsiderAdvantage, in the field October 21 and 22, had Hinson ahead 47 to 45.¹⁴ A review of the 12 polls preferred by The New York Times found seven for Hinson and five for Turek, which is another way of saying the race sits inside the noise.³¹ Silver Bulletin moved Iowa to lean Democratic on October 21, making it the only model in the table that now favors Turek.¹ When a model breaks from a market this late, one of them is reading the same evidence differently, and here the evidence is a Republican state with a Democratic money advantage and a Republican cash advantage.

Nebraska is the cheapest fourth seat on the board and the race where the average and the forecast disagree most. Call the Map’s average of 14 polls has Ricketts ahead by 0.3 points, and its partisan adjustment flips the race to Osborn by the same margin, which is a polite way of saying nobody knows.³² The two most recent polls disagree in the usual direction: a GBAO survey for Osborn’s side had him ahead 46 to 45, and an Impact Research poll for the Democratic side had Ricketts ahead 50 to 48.³² Osborn’s independence is the whole of his argument and the whole of his problem. A senator who will not caucus with either party cannot promise a committee assignment, and a state the president carried easily does not usually send one to Washington to find out.

Kansas is the margin seat, and it now has the spending to prove it. Both parties are now paying to contest a state that has not sent a Democrat to the Senate since 1932: about $71 million in outside money, a Republican incumbent who raised $905,000 since the primary against a challenger who raised $8 million, and a president who came to Park City to say so.³ Kalshi last published a Kansas price on October 8, at 62 percent for Marshall, before the debate and before the spending figures became public.³³ The honest caution is unchanged. Kansas produced competitive September polls in 2014 and 2020 and elected Republicans by double digits in both Novembers, and a state with a 26 percent Democratic registration does not flip because one quarter’s fundraising report was good.³⁴

Florida is not a race, but it is a six-point state in a year when the president’s approval is in the thirties. The Orlando Chronicle’s average of the three polls that meet its quality bar has Moody ahead 49.3 to 43.0, including a University of North Florida survey at 49 to 45.³⁵ Inside Elections moved Florida from solid to likely Republican on October 15, which is the correct rating for a seat that is not competitive but is no longer safe.

Louisiana has become the most interesting long shot in the country. Two polls paid for by Jamie Davis’s campaign, conducted by different firms, put him four points behind Julia Letlow, and he out-raised her nearly three to one in the third quarter, with more than 90 percent of his money from individuals.³⁶ Letlow’s name recognition is 82 percent against his 41, but only 28 percent of independents rate her positively, and the pollster John Couvillon told The Lens that he is seeing Republican defections “in about the 12% range.”³⁶ On the first day of early voting, Black voters were 32 percent of the electorate, a share higher than in nearly any election of the past 18 years.³⁶ A four-point race with nine days left and that electorate turning out is a race.

Mississippi is the state that explains why the map feels wider than the math. Cook moved it from solid to likely Republican, and the Senate Majority PAC put $1.2 million into television and digital advertising through Election Day in a state the president carried by 23 points, where no Democrat has won a Senate seat since 1982.⁸ A PAC spokesperson said, “Colom is running a competitive race that has put Hyde-Smith on her heels.”⁸ The rating says the seat is not close. The spending says the party believes it might be.

Montana has a three-way race and an independent with a real base. A Quantus Insights survey released October 20 had Kurt Alme at 44, Seth Bodnar at 28 and Alani Bankhead at 21, with a margin of error of four points.³⁷ The independent Bodnar is the reason the seat is not entirely safe, and also the reason it remains Republican-favored: 28 percent in a four-way race is not a path to 50.

South Carolina is the newest entry in the competitive tier and the least comfortable for the Republicans. PollingSource’s average of three polls has Darline Graham ahead 47.3 to 45.3, with a Democratic internal showing Andrews ahead by one and two Republican polls showing Graham ahead by three and four.³⁸ Cook’s move to lean Republican is the right response to a two-point average and a race that has tightened all month. It is not a prediction that the seat flips. It is a prediction that the final week has to be spent there.

The nine safe Republican seats, Alabama, Arkansas, Idaho, Kentucky, Oklahoma, South Dakota, Tennessee, West Virginia and Wyoming, are not in play.

The Math

Sabato’s Crystal Ball has not published a new rating since October 14, and its count still stands at 49 seats at least leaning Democratic, 46 at least leaning Republican and five toss-ups: Kansas, Iowa, Maine, Ohio and Texas.¹ That arithmetic has not changed in a week, but the ratings underneath it have. Cook’s Minnesota move takes a Democratic seat down a notch, and its South Carolina and Mississippi moves take two Republican seats down a notch, which is what a map looks like when it widens at the edges while the center holds. The requirement that follows is still stark: Republicans need four of the five toss-ups for a 50-50 chamber with the vice president’s tiebreaker, and Democrats need two.¹

The market has moved further than the raters. Kalshi’s 68 percent for Democratic control is a six-point gain in eight days.² Its seat distribution puts the most likely single outcome above 52 seats at about 38 percent, with exactly 52 at 22 percent, 51 at 11 percent and 50 at 13 percent.² Read plainly, the market’s favorite is a 52-seat Democratic caucus, and its second favorite is a 50-50 Senate that the vice president controls. That is the same thing as saying the market sees a real chance that one seat decides the chamber, which is why the Kansas price and the Alaska lawsuit matter more than their single seats.

The paths have not changed in their order, only in their price. The market path is North Carolina, Ohio and Texas plus one of Maine, Alaska, Iowa or Nebraska. The consensus path is North Carolina, Ohio, Maine and Alaska, which requires winning the top four and assumes the defense holds everywhere. The nightmare is any loss among Michigan, Minnesota or New Hampshire, because every defensive loss raises the requirement to five flips, and five flips means winning Kansas or Nebraska or Florida on top of everything else.

The Kansas connection is the argument in miniature. A month ago Kansas was a curiosity, two weeks ago it earned a row in the table below, and this week it became the most expensive Senate race in its history: $71 million in outside spending, a Republican incumbent out-raised nine to one since the primary, a presidential rally, and a debate in which both men claimed the state was the fulcrum of the majority.³⁴ What has not happened is the other half. The market still prices Marshall at 62 percent, and that price is two weeks old.³³ If the Democrats sweep North Carolina, Ohio and Texas, Kansas is not needed. If any one of those slips, Kansas becomes the margin seat, and a margin seat that the market prices Republican is a margin seat that stays where it is.

The ratings table

The columns are refreshed on different schedules and the differences are real, so the dates are given per column rather than averaged: Cook October 20; Inside Elections October 15; Sabato October 14; RealClearPolitics October 12; Decision Desk HQ October 22; Silver Bulletin October 21.¹ Three rows moved since last week. Minnesota moved toward the Republicans in Cook’s column, South Carolina moved toward the Democrats in Cook’s, and Iowa moved toward the Democrats in Silver’s. Mississippi moved toward the Democrats in Cook’s rating as well, but it is not in this table because no forecaster has it closer than likely Republican.

A word on what the columns mean. PVI is the Cook Partisan Voting Index, and it is the only backward-looking number here. It measures how far a state leans from the nation as a whole, computed from the 2020 and 2024 presidential two-party vote shares weighted toward 2024, and Cook says of it that the race ratings reflect an outlook for which party will win the next election, while the PVI takes a longer view. It earns its place as contrast, which is what makes “Tossup in an R+6 state” legible, and not as a forecast. Lean is a strong prior across the whole map: Pew found 158 of 176 Senate elections since 2012, 89.8 percent, went the way of the state’s presidential carry. But applied as the sole predictor to the races this series tracks, lean calls five correctly and misses four, Georgia, North Carolina, Ohio and Texas, every one of which it would call Republican and every one of which the forecasters now favor for the Democrat. That is not a contradiction but a selection effect: a race earns its way onto this table precisely by being a case where the baseline does not settle it. Never argue from PVI against the forecasters, because where the two disagree, the forecasters are the point of the exercise.

Two traps make that column look inconsistent with numbers quoted elsewhere. The first is that PVI is denominated in vote share for one party, not margin, which makes it roughly half the size of a margin-style lean: Cook calls Texas R+6 while Silver’s partisan lean calls the same state R+11, and both are right, because the denominators differ. The second is that the scale flattens elasticity, how hard a state swings when the nation swings, and elasticity is orthogonal to lean, correlating with it at only r = +0.12 across these races. New Hampshire, at 1.15, and Maine, at 1.13, are among the most elastic states in the country; Georgia, at 0.90, is among the most inelastic, which is why it can be rated comfortably Democratic while leaning Republican on the baseline. A state can be strongly partisan and highly elastic at once, which is why a lean column can never stand in for a competitiveness column.

IE is Inside Elections, Nathan Gonzales’s nonpartisan newsletter, one forecaster among several. The rating scale runs, strongest to weakest, solid, likely, lean, tilt, and toss-up, and Tilt is Inside Elections’ own rating: Cook, Sabato and Decision Desk HQ do not use it, so a Tilt appearing only in the IE column is not a disagreement but a finer scale. Decision Desk HQ publishes win probabilities that map onto the same ladder: above 95 percent is solid, 80 to 95 likely, 65 to 80 lean, and 35 to 65 a toss-up for either party. (flip) marks a rating that forecasts a change of party from the current holder.

A note on the sources. One column in that table is arithmetic and six are not. PVI alone can be reproduced from certified FEC totals, which we have done, matching Cook’s published figures; everything else is somebody’s assessment. Cook, Inside Elections, Sabato and RealClearPolitics are analyst judgment, which means their ratings cannot be independently reproduced from any underlying data, because the rating is the data. That is expert opinion in a table rather than a measurement, and it is not a scandal: election analysis is not a solved problem. RealClearPolitics is worse than unaudited, because it blocks automated reading, so we cannot say whether that column is independent judgment or a composite of the others; treat it as unlisted rather than as a seventh opinion. Decision Desk HQ and Silver are published models, and Decision Desk HQ is the most independent column here, using neither expert ratings nor Cook’s PVI and deriving its prior from the most recent presidential margin adjusted for the national swing. Silver’s FLIPR is not independent in the same way, because in its default build it feeds the Cook, Inside Elections and Sabato ratings in as an input at roughly one-sixth weight, so where his column agrees with theirs the agreement is partly built in and is not corroboration. All six read the same polls and talk to the same operatives, which means six agreeing columns are not six confirmations; the signal is where they disagree, and this week that is Iowa, Kansas, Minnesota and Nebraska. Freshness differs per column and is stated above rather than averaged. Sabato’s column is now six days staler than Cook’s, which is why the market has moved further than the table.

The Close

Nine days is the part of a campaign when the arguments stop changing and the arithmetic starts to move. This week the arithmetic moved in three places at once: a Republican incumbent in Kansas out-raised nine to one, a Republican nominee in North Carolina facing a party that has retreated from the state, and a Justice Department suing to keep a Senate seat from being certified before a single ballot is counted. None of those is a result. All three describe a party that has stopped expecting the map to hold and started deciding which parts of it to save.

The country is not being asked, in this last week, whether it likes the president. That question was answered months ago, in the thirties and staying there. It is being asked whether a midterm electorate that is angrier about prices than about anything else will spend its anger on the party that holds the Senate, and whether the states that decide the chamber, Kansas and Iowa and Maine and Alaska, will follow the national mood or their own older gravity. Kansas has voted Republican for ninety-four years. Maine has re-elected Susan Collins after the polls said she would lose.³⁴ Alaska counts its ballots in rounds, and a race this close may not be settled until the last one. Nine days is enough time for any of those habits to reassert itself, and enough time for all of them to break at once.

Notes

PRH | huffmanwrites.org | © Philip Huffman