AFOS Analytics
Real-time political pricing
Prediction markets × polls × news
no smoothed averages
United StatesThe previous issue opened by saying that the contract on Democratic control of the House had crossed the week with a single distinct value across 46 readings. This week it moved 2.00pp, from 92.50% at the close of 2 October to 90.50% at the close of 9 October, and it moved in steps: 92.50% through 6 October, 91.50% on 7 October, 90.50% on 8 October, and flat there in the final session. What makes the move relevant is not its size, it is that it survives normalisation. The sum of the pair came to exactly 100.00% at both edges of the window, so the implied probability fell the same 2.00pp that the price fell, and there is no book spread to discount. That is the difference between this week and the previous one, in which the sum of the pair reached 101.00% midway and the implied figure for Democratic control of the House fell to 91.58% without the price moving a single tick. In the Senate the reading is the opposite and it requires both measures. The contract on Democratic control closed the week at 62.50%, 2.00pp below the 64.50% of 2 October, but the sum of the pair OPENED over the period, from 100.00% to 99.00%, and normalised the retreat is 1.37pp, not 2.00pp. The Republican side of the same book shows the mirror image: it rose 1.00pp raw, from 35.50% to 36.50%, and 1.37pp normalised. Two points raw and one point thirty seven normalised are not the same number, and the difference between them is the level of overpricing the book carried at one edge and did not carry at the other. And it was inside this week that the Senate Democratic contract touched 67.50% on 4 October at 23:30 UTC, which is the top of the entire series recorded since 14 April, before giving back five points by Friday's close.
1.Executive Summary
The three figures above are of different natures and do not combine. The first two are contract prices, read live and confirmed by a double reading separated by 8 minutes, with a tolerance of 0.20pp. The third is an arithmetic average of published polls, on a 30 day rolling window. This issue measures the behaviour of each one over the week of 5 to 9 October, with the delta anchored at the close of the previous Friday, 2 October, and it refuses the subtraction between them.
2.Why this brief does not subtract the market from the poll
The question comes up in every trading desk conversation, and this week it comes up with new figures: if the market pays 90.50% for Democratic control of the House and the polling average gives a Democratic lead of 7.57 points, what is the distance between the two readings? The answer is still that the question has no numerical answer, and the reason is one of units.
Subtracting 7.57 from 90.50 would produce the number 82.93, which is not a probability, is not a vote point and would change size if the poll were expressed on another scale, for instance as a share of the two party vote. The 2012 case is the short historical example: the Democrats won more votes for the House and fewer seats, because converting vote into seats depends on geography and on district design. This brief reports the two readings side by side and never the difference between them.
There is a better comparison available this week, and it involves no subtraction: the market publishes TWO books on the same House question, a binary on control and a distribution of seat ranges, and those two are comparable with each other because they measure the same magnitude. That is what the scenarios section below does, and the result is that they do not contradict each other.
3.Weighted scenarios for the week
The scenarios below are not narrative: they are read in the prices of the contracts the market already publishes. In this issue they use the two control binaries and the distribution of Republican House seat ranges, whose sum closed the week at 103.10% and stands at 104.25% at the certified reading, inside the 95% to 105% gate at both ends. The ranges in that distribution are cited normalised by the sum of the book, because the partition is clean across the ten ranges and the excess is spread approximately uniformly among them, which preserves the relative shape.
The binary on Democratic control of the House closed the week at 90.50%. The seat distribution says the same thing by another route: the ranges that give the Republicans 214 seats or fewer sum to 86.28% normalised, and a majority in the House requires 218. Adding the range of 215 to 219 seats, which straddles the threshold and cannot be assigned to either side, the ceiling reaches 92.56%. The binary's 90.50% falls inside that interval of 86.28% to 92.56%, and that is what can be asserted: two different books, on the same question, with no contradiction between them.
The binary on Republican control of the House closed at 9.50%, against 7.50% on 2 October, and the path shows the intraday floor of 6.50% on 3 October at 02:00 UTC before the rise. By the seat distribution, the ranges strictly above 219 sum to 7.43% normalised, and adding the range that straddles 218 the interval reaches 13.72%. The binary's 9.50% also falls inside that interval. The single most expensive range in the whole book remains fewer than 190 Republican seats, at 26.38% normalised, and the second is 195 to 199, at 13.91%.
The contract on the election happening as scheduled closed the week at 98.70%, 0.50pp above the 98.20% of 2 October, which leaves the complementary outcome at 1.30%. Inside the window it touched 98.90% on 9 October at 11:30 UTC, which is the top of the entire series recorded since 29 July, and the intraday floor was 98.00% on 4 October at 23:00 UTC, an amplitude of 0.90pp across 44 readings. It is the quietest of the five books this brief follows and the smallest by volume, with USD 0.63M accumulated at the certified reading of 11 October.
4.Indicator Grid
| Contract | Current | Δ week | Vol USD acc. | Implied reading |
|---|---|---|---|---|
| House, Democratic control | 90.50% | -2.00pp | USD 9.18M | A stepwise move, 91.50% on 7 October and 90.50% on 8 October, held on Friday. Intraday top of 93.50% on 3 October at 02:00 UTC and floor of 90.50% first reached on 8 October at 05:00 UTC. Pair at 100.00% at both edges: raw and normalised say the same thing. |
| House, Republican control | 9.50% | +2.00pp | USD 7.25M | Intraday floor of 6.50% on 3 October at 02:00 UTC, top of 9.50% on 8 October at 07:00 UTC, amplitude of 3.00pp. New money across the whole House book over the period: USD 2.63M, four times that of the Senate book. |
| Senate, Democratic control | 62.50% | -2.00pp raw, -1.37pp normalised | USD 3.68M | Top of 67.50% on 4 October at 23:30 UTC, which is the maximum of the entire series since 14 April, and a fall of five points to Friday's close. The sum of the pair opened from 100.00% to 99.00% over the period, and that is why the two measures disagree by 0.63pp. |
| Senate, Republican control | 36.50% | +1.00pp raw, +1.37pp normalised | USD 2.64M | Here the normalised figure moves MORE than the raw one, which is the mirror of the Democratic side. Intraday floor of 33.50% on 4 October at 23:30 UTC, at the same instant as the top on the other side, and a close at 36.50%, the top of the window. |
| Election happens as scheduled | 98.70% | +0.50pp | USD 0.63M | Top of the entire series at 98.90%, reached on 9 October at 11:30 UTC. The path inside the window is not monotonic: 98.00% on 6 October, 98.50% on 7, 98.40% on 8 and 98.70% on Friday. |
| Republican House seats, distribution across 10 ranges | sum 103.10% at the close | from 106.90% on 2 October | USD 0.32M | It entered the 95% to 105% gate over the course of the week and remains inside at the certified reading, at 104.25%. Clean partition across the ten ranges, with no overlap and no hole. It is the smallest book by volume in this table and the only one that gives the seat dimension, which neither binary touches. |
5.Liquidity and market structure
6.Electoral calendar through November
| Date | Sample | Why it matters | |
|---|---|---|---|
| 3/Nov, Tuesday | Midterm election day | 435 House seats and one third of the Senate | It is the only date on this grid that settles the contracts. Every price cited in this issue is the implied probability of an outcome that becomes known on that day. |
| 5 to 9/Oct | Early voting opened in several states over the course of the week | state calendar, date by date | It is a calendar marker and not a price event. No contract move in this issue is attributed to it, because the time window differs by state and was not established against the series. |
| 8/Oct, Thursday | Trump says he will not resume strikes on Iran before the elections | statement by the president | The contract on the election happening as scheduled CLOSED that day at 98.40%, 0.10pp below the previous day's close, and the top of the series came the following day. The series does not support attributing the contract's move to this statement, and so the brief does not attribute it. |
| 9/Oct, Friday | Cook shifts 18 House races toward the Democrats | 18 districts reclassified | Over the same interval, the price of Democratic control of the House FELL 2.00pp. The fall occurred on 7 and 8 October, before this publication, so the order of events supports no causal relation in either direction. These are two readings of the same question pointing in opposite directions in the same week. |
| 6/Oct, Tuesday | National Republican groups move into Alaska, and a super PAC puts USD 5M into South Carolina | 2 statewide Senate races | The Senate control contract is national and binary; the coverage beneath it is statewide. This brief does not convert one reading into the other, and records only that new money in the national Senate book over the period was USD 0.65M, a quarter of the House figure. |
The grid above has one date that settles a contract, 3 November, and four that settle none. The four are included because they are verifiable and dated, and they come with the measurement alongside when one exists. When the series does not support the link between the fact and the price, as in the cases of 8 and 9 October, the brief says so and stops there.
7.Watch list, week triggers
- The 215 to 219 seat range, at 6.28% normalised It is the only range in the seat book that straddles the 218 threshold, and for that reason it is the width of the interval within which the two House books can agree without contradicting each other. If it gets more expensive, the interval widens and the comparison between binary and distribution loses resolution; if it gets cheaper, the interval closes and the two books begin to hold each other to account.
- The sum of the Senate pair, at 99.00% at the close of the week It stood at 100.00% on 2 October and has already returned to 100.00% at the certified reading of 11 October, with the Democratic side back at 63.50%. As long as that sum oscillates between 99% and 100%, every Senate delta needs both measures, because a full point of raw can be worth 1.37pp or 0.37pp of implied probability depending on the edge.
- The leading edge of the generic ballot window, stalled at 5 October The average served uses a 30 day rolling window, and the five aggregators that publish a running average declare field dates up to that same 5 October. With no new round entering, the base of the average falls from 41 to 16 rounds by 22 October through the edge effect, and the number that changes first is the sample size, not the level.
- The prediction market dispute at the Supreme Court During the week the American football league asked the Supreme Court to take up the case on event contracts. It is a structural fact about the venue where the prices in this brief are formed, and not about any of the races those prices cover.
8.Methodology
Prices read on the AFOS's own route with the country parameter and an uncached read, and confirmed by the capture lock: two readings separated by 8 minutes, released only if they agree within 0.20pp, with the two required timestamps distinct. The certified reading for this issue is 11 October 2026 at 21:25:28 UTC. The series for each contract is read in the database backup and never on the history route, because that route caps the window at 90 days: on the Senate contract that would hide 162 points and would make the real floor of the series, 39.50%, appear as 43.50%. Closes are the last point recorded on each day, and tops and floors are intraday, because a daily close hides extremes. The weekly delta is anchored at the close of the previous Friday, 2 October, which is the convention of this United States series. Any range distribution is cited only when the sum of the book falls between 95% and 105%, and it is cited normalised by the sum. The generic ballot average is a simple arithmetic one, unweighted, on a 30 day rolling window, with one round per pollster per wave by the hierarchy likely voter, registered voter, adults, and with a wave published under a different instrument kept out of the count and declared with the pollster's own wording. Book depth is not cited in this brief, by an editorial decision of May 2026.
9.Additional reading · macro coverage
The items below are the outlets cited in the sections above, with the address of the article itself. Items the collection reached only through the aggregator were dropped from this list, because the aggregator's address does not point to the outlet.
- The Washington Post, 9/Oct · Cook shifts 18 House races toward the Democrats, www.washingtonpost.com/politics/2026/10/09/cook-moves-18-more-house-races-toward-left/
- CBS News, 8/Oct · Trump promises not to resume strikes on Iran before the elections, www.cbsnews.com/live-updates/iran-war-nuclear-donald-trump-vance-rubio-strait-of-hormuz/
- Axios, 8/Oct · The same statement, framed as the United States not attacking before the vote, www.axios.com/2026/10/08/trump-iran-strikes-midterm-elections
- NewsNation, 8/Oct · The American football league asks the Supreme Court to take up the prediction market case, www.newsnationnow.com/us-news/sports/nfl-supreme-court-predition-market-kalshi/
- The Washington Post, 7/Oct · The extraordinary amount of money pouring into the midterm election, www.washingtonpost.com/politics/2026/10/07/extraordinary-amount-money-pouring-into-midterm-elections/
- The Washington Post, 6/Oct · National Republican groups move into the Alaska Senate race, www.washingtonpost.com/elections/2026/10/06/gop-sen-dan-sullivan-gets-major-national-help-tough-reelection-bid/
- CBS News, 6/Oct · Republican super PAC puts USD 5 million into the South Carolina Senate race, www.cbsnews.com/news/south-carolina-senate-race-fundraising-darline-graham-annie-andrews/
- The New York Times, 9/Oct · A Democratic group pumps millions into the Kansas Senate race, the same one the previous issue recorded as receiving displaced Republican spending, www.nytimes.com/2026/10/09/us/politics/adam-hamilton-roger-marshall-kansas-senate.html
- The Hill, 8/Oct · Trump campaigns in Alaska in the sequence of events over the period, thehill.com/homenews/administration/6136988-trump-wasilla-alaska-rally/
Get it by email
AFOS Tradeoff is the weekly read on Brazilian political risk, for people who decide on data rather than headlines.
No spam. Unsubscribe anytime.
After signing up you choose whether to receive it in English, Portuguese or Spanish.