AFOS Weekly · USA-2026 midterms Political Risk Weekly

AFOS Analytics

Real-time political pricing
Prediction market × polls × press
no smoothed averages

Harvard Dataverse · DOI 10.7910/DVN/XRUT8U
United States·Issue №10·Week of 2026-10-01 to 2026-10-08·published thursday 07:00 BRT

What the prediction market did

The five published prices hold USD 21.78M and took USD 3.00M of new money, four fifths of it in the two House books. The chambers did not move together: House Democrats lost 2.00pp, Senate Democrats gained 1.00pp, on different days.

House · Democrats

90.50%

↓ 2.00pp on the week

On USD 8.59M, with USD 1.32M of new money, the most of any book. It held 92.50% from October 1 to October 6, touched 93.50% on October 3, then closed 91.50% and 90.50%: the whole move in the last two days, across 48 readings.

House · Republicans

9.50%

↑ 2.00pp on the week

On USD 6.55M, with USD 1.05M of new money. The mirror: 6.50% on October 3, 9.50% on October 8. The pair summed to 100.00% at both ends, so the 2.00pp survives normalisation whole.

Senate · Democrats

63.50%

↑ 1.00pp on the week

On USD 3.51M, with USD 0.31M of new money. It opened at 62.50%, added two points on October 2, reached 67.50% late on October 4, then closed lower four days running: 65.50%, 64.50%, 64.50%, 63.50%.

Senate · Republicans

35.50%

↓ 1.00pp on the week

On USD 2.52M, the thinnest control book, with USD 0.20M of new money. It reached 33.50% on October 4, the lowest of the series. This pair sums to 99.00%, so the Democratic gain reads 1.01pp normalised.

The fifth price, on the election happening on schedule, closed at 98.50%, up 0.40pp, on USD 0.62M. That is its highest since the series began on July 29, first reached late on October 7, and its whole weekly range is 0.50pp.

What the polls did

The spread comes before the average, because the spread is the finding. Six rounds with fieldwork inside the week entered the average, and they disagree by five points.

Widest

Cygnal and CBS News/YouGov · D+9

Narrowest

ActiVote · D+4

Spread between institutes

5 points

Three of the six closed fieldwork on the same day, October 2, five points apart: ActiVote at D+4 on 1,000 likely voters with a margin of 4.4, the widest of the week, Cygnal at D+9 on 1,500 with a margin of 2.53, and CBS News/YouGov at D+9 on 1,572. The two largest samples both landed at D+7: Morning Consult with 17,659 and Reuters/Ipsos with 3,526.

What the press reported

What the fixed-list outlets covered, October 1 to October 8

Republican money left North Carolina, and four outlets reported it the same day. On October 2 the New York Times ran "As Prospects Dim, G.O.P. Pulls Money Out of North Carolina Senate Race", Politico reported top Republican groups retreating, NBC News reported a super PAC pausing its spending there and moving the money to Kansas, and the Guardian covered the withdrawal. Politico reported the same day that Kansas had become a surprise Senate battleground.

The money went elsewhere. On October 2 the Hill reported Vance rallying in Alaska. On October 6 the Washington Post reported national Republican groups pouring into Alaska and CBS News reported a super PAC putting five million dollars into South Carolina. On October 8 the Hill reported Trump heading to Alaska himself.

A forecaster moved the same way: on October 6 the Washington Examiner reported Sabato's Crystal Ball shifting Alaska and South Carolina toward the Democrats, and on October 7 the Washington Post opinion page wrote that polling in Kansas had shocked Republicans.

Underneath it, early voting opened, with state dates in USA Today on October 5 and queues before dawn in Ohio reported by Reuters on October 6; and on October 8 CBS News, the Guardian and Axios reported Trump promising not to strike Iran before the election.

Our measurement: A forecaster rates individual races, the contract prices control of a chamber: different objects, and we are not calling them a contradiction. But in the four days after the North Carolina withdrawal was reported, the Senate price for Democrats ran 67.50%, 65.50%, 64.50%, 64.50%, 63.50%.

The crossing of the week

The peak came before the news that pointed the same way

The Senate book added two points on October 2, the day four outlets reported Republican money leaving North Carolina, and reached its series high of 67.50% two days later. It then fell on October 5, 6, 7 and 8, including the day a forecaster moved two more Senate races toward the Democrats. We claim no causation in either direction. The sequence is the fact: the high sits behind the good news, not after it.

Two chambers, one election, opposite directions

House Democrats fell 2.00pp and Senate Democrats rose 1.00pp in the same eight days. The two contracts are not two readings of one thing: the whole House runs across 435 districts every two years, a third of the Senate on a map fixed six years ago. Every Senate story this week was North Carolina, Kansas, Alaska or South Carolina, and none of those is a statement about 435 districts.

How to read this number

The biggest polling number of the week is the one we leave out

The Economist/YouGov closed fieldwork on October 5 and put Democrats ahead by 12 among 995 likely voters, the widest margin published this week. It is not in our average, and has not been since September 8. The reason is in the pollster's own tab report: that question is asked using the names of the candidates running in the respondent's district. A generic ballot asks which party you would back for Congress; a named ballot asks about two specific people. Different questions, and the named one tends to run wider, because a voter unsure about a party can still like a neighbour. Our average answers one question, so it takes only the rounds that asked it. The exclusion is recorded with the pollster's own wording in the published data file, and the rounds return the day the house does.

Sources

Prediction market, October 8 closes

Polling, primary sources

Press cited in sections 4 and 5

Method

Market prices in this issue are October 8 closes, the last point recorded that day in our daily database backup, compared with the October 1 closes that Issue No. 9 published. Both ends of each binary pair are printed raw and normalised, because a pair that does not sum to 100 moves one side without the other side trading. The polling average is a simple, unweighted mean of one round per polling firm per wave over a rolling thirty-day window, with the hierarchy likely voters over registered voters over adults when a firm publishes more than one. Rounds that asked a named-candidate ballot instead of the generic ballot are excluded by wave, with the pollster's own wording recorded in the published data file. We never subtract a market price from a polling margin: one is a probability of controlling a chamber and the other is a margin in vote share, and in 2012 Democrats won more House votes and fewer seats.

AFOS
AFOS Analytics
Global Political Risk Intelligence · GLOBAL BY DESIGN
www.afos-analytics.com · contact@afos-analytics.com · github.com/AFOS-Analytics

Get it by email

AFOS Weekly follows the US midterms every Thursday, written for voters: where the prediction market, the polls and the press disagree, and what each one measures.

No spam. Unsubscribe anytime.

After signing up you choose whether to receive it in English, Portuguese or Spanish.

AFOS Weekly · Issue No. 10 · The Senate price hit its series high on Sunday and gave back four points | AFOS Analytics