The chance of a Democratic Senate rose from 69% to 76% after a day of new state polls. Iowa tipped to Josh Turek by 0.2 points, and the New Hampshire and Vermont governor races moved sharply after one poll.

Democrats' Senate chance edged from 70% to 69% and their House chance from 93% to 92%, as the generic ballot average fell 0.4 points. Florida and Ohio shifted slightly on new polls.

Adam Hamilton now leads Roger Marshall in Kansas, with a 55% chance on the forecast. The Senate chance for Democrats is 65%, up from 62% yesterday, while the House chance is 91%, down from 93%.

Barry Moore's Senate lead fell from 20.5 to 13.8 points and Tommy Tuberville's from 14.5 to 8.0, each after one Blueprint Polling survey. Both Republicans remain clear favourites.

A New York Times/Siena poll lifted Peggy Flanagan's Minnesota Senate lead from 6.2 to 8.4 points. Maine and Nebraska governor averages moved the other way, and the Democratic Senate chance rose a point to 61%.

Senate and House odds were unchanged. Two governor's races moved in opposite directions: Michigan towards Jocelyn Benson after a new YouGov poll, Nevada towards Joe Lombardo with no change in the poll count.

Tom Begich's lead in the Alaska governor race grew from 1.4 to 1.8 points, the largest move of the day. Most other averages slipped slightly towards Republicans, and the Democratic Senate chance eased from 61% to 60%.

New York Times/Siena polls moved Pennsylvania's 1st towards Robert Harvie, who now has a 72% chance, and Pennsylvania's 10th towards Scott Perry, leaving Janelle Stelson at 51%. The national odds did not change.

Ashley Hinson now leads Josh Turek in Iowa after a new InsiderAdvantage poll, and Democrats' Senate chance eased from 63% to 61%. Virginia's 1st House district also tightened sharply.

Vermont's governor race went from a Democratic lead of 8.4 points to a Republican lead of 3.0, and Oregon's tightened to 2.4. Both averages now hold more polls than a day ago. The Senate edged down to 63% for Democrats.

The published chances of a change in Senate control run from below even to three in four. Most of that gap is not about the polls. It comes from where the chamber sits relative to the line, and from a handful of modelling choices nobody has settled.

Eleven polls from Wedgewood, which admitted on 21 October that its surveys were fabricated, were sitting in our averages. Ten races barely noticed. One moved five points, and the reason is the useful part.

Roger Marshall's Kansas lead shrank from 2.7 points to 0.3 and his chance fell from 68% to 52%. Texas moved towards Talarico, and Democrats' Senate chance rose from 62% to 64%.

We looked at 11,500 Senate, governor and House polls from 1998 to 2022. The last fortnight sharpens the picture a little, not a lot, and a lead under two points this far out has historically been a coin toss.

A new poll put the Alaska Senate race at a single point, and the forecast moved it to a dead heat. In Nebraska, a poll for a Democratic sponsor showing a one-point race moved the forecast towards the Republican. Both follow from how the model reads a poll.

New polls narrowed both Ohio races and the Florida governor's race moved the other way. The Senate forecast fell four points, and about two-thirds of the fall in one of its national inputs turned out to be a flaw in the method, fixed that night.

Three changes went in on 19 October, all of them following documented practice from other forecasters. Together they moved the Senate forecast from 58% to 63%. The polls did not move that much.

Several outlets led with worrying economic data and an unhappy electorate. The forecast moved by less than half a point, and our own research on what moves the congressional vote explains why.

The Iowa Senate race moved to within half a point. Six of the seven closest Senate races are now in states Donald Trump carried in 2024, which is where both parties are spending their final weeks.

The day's coverage was about money. The forecast's biggest move was in Kansas, where four polls say the race is level and the state's history says it is not.

We tested twenty economic measures against 2,130 polls over twenty months. The economy reaches presidential approval clearly and stops there, and the reason is more interesting than the result.