Every Monday, we pull the five most-traded questions on Kalshi and Polymarket that actually settle inside the week, research them properly, and then make a prediction while including all our sources.
Each prediction gets a Geek Score from 0 to 100 depending on our level of certainty. Here’s the board for this week.
Jobs numbers this month?
Kalshi · Resolves Friday, September 4
Our pick: Above 50,000
Price at 2:40 p.m. ET, Aug 31: Yes 49c – No 52c
Our call: NO, Geek Score 30
On August 28 the BLS published its preliminary benchmark revision and removed 79,000 jobs from the year through March 2026, with retail trade alone losing 154,600. That landed four days before this market has to settle, and the crowd still has it close to a coin flip at 49c.
The recent record is worse than the pricing suggests. July payrolls fell 23,000 against a Dow Jones consensus of plus 83,000, as CNBC reported on August 7. June was revised down to plus 20,000 and May to 63,000, which drags the 12-month average to 34,000 a month. Only one of the last three prints cleared 50,000, and that was the revised May figure.
Underneath the headline, private payrolls added 30,000 in July while government shed 53,000, driven by a 50,000 drop in local government education. Wage data agrees: average hourly earnings rose two cents, and 12-month growth of 3.2% is the slowest since May 2021. Forecasters still expect around plus 80,000 for August, per VT Markets, which is roughly what they expected in July.
The risk: the July collapse was concentrated in one line item, and local government education is the most seasonally awkward series in the report. Those distortions frequently reverse the following month. A normal private-sector print of 40,000 to 60,000 with a smaller government drag clears this threshold without anything improving, which is why the score is 30 rather than 15.
Unemployment in August
Kalshi · Resolves Friday, September 4
Our pick: Above 4.1%
Price at 2:40 p.m. ET, Aug 31: Yes 60c – No 40c
Our call: NO, Geek Score 35
Labour force participation fell to 61.4% in July, the lowest reading in more than five years. That number does more work here than any hiring figure.
Calling NO on payrolls above 50,000 and NO on unemployment above 4.1% in the same week looks contradictory until you separate the two surveys. Payrolls count jobs; the jobless rate counts people who say they are looking. The rate fell to 4.1% from 4.2% in July with 6.9 million unemployed, and it has come down from 4.4% last September over exactly the stretch in which payroll growth collapsed. People are leaving the labour force faster than employers are cutting, and a shrinking denominator holds the rate down even in a weak market. Consensus for August is 4.2%, so the crowd at 60c is paying for a return to the trend that participation keeps interrupting.
The risk: the household survey is small and noisy, and 0.1 point moves happen for no reason at all. If the unrounded July rate sat near 4.14%, almost any upward drift prints 4.2% and this call is dead on rounding alone. That fragility is why the score sits at the top of the NO range rather than deep inside it.
Colorado vs. Georgia Tech
Polymarket · Resolves Thursday, September 3, 8:00 p.m. ET
Our pick: Colorado (moneyline)
Price at 2:40 p.m. ET, Aug 31: Yes 33c – No 68c
Our call: NO, Geek Score 25
Bobby Dodd Stadium on a Thursday night in prime time on ESPN, with Georgia Tech laying 6.5 points. The same two teams met last season in Boulder, where Georgia Tech won 27-20 as a 3.5-point road favourite. The line has moved 10 points in Georgia Tech’s favour since, which is most of the story.
Colorado finished 2025 with three wins and rebuilt the defence over the offseason. Both sides start a new quarterback: Georgia Tech brought in Alberto Mendoza from Indiana after Haynes King left for the NFL. At 33c the market gives Colorado a real chance, and our 25 is only eight points below it, so this section is closer to agreement than we usually publish. The related legs price consistently, with Colorado +6.5 at 52c and the total at 50.5, on $9.76K of volume, the platform’s busiest college football market this week.
The risk: two first-time starters in a season opener usually produces a slower, sloppier, closer game than a 6.5-point line implies, and openers are where roster turnover is least visible in the numbers. Deion Sanders has had a full offseason to fix the unit that lost this exact game a year ago.
WILDCHILD: First Week Album Equivalent Units
Kalshi · Resolves Sunday, September 6
Our pick: Above 70K
Price at 2:40 p.m. ET, Aug 31: Yes 65c – No 40c
Our call: YES, Geek Score 70
This leg repriced hard today, gaining 28 points, while the 60K leg gained 21. Something moved the room, and we could not source what.
The case for the album is the catalogue behind it. The artist’s debut, You’ll Be Alright, Kid, opened at No. 5 on the Billboard 200 and went gold within a day of release. Ordinary was the top-selling song of 2025, held No. 1 on the Hot 100 for 10 weeks and spent 36 weeks atop Adult Contemporary. Total catalogue streams are above 11 billion. WILDCHILD arrived August 28 on Atlantic with four pre-release singles already out: Fever Dream, Fine Place to Die, Passenger and Rescuer. For a pop act operating at that scale, 70,000 equivalent units in week one is a low bar.
The risk: 13 tracks and 38 minutes 55 seconds is a short record, and streaming equivalent units scale with track count. Forbes noted in July that none of the singles had matched Ordinary, so the launch lacks a breakout. Our 70 sits five points above the market, which is not real edge, and the unexplained 28-point move means the crowd may be trading on information this call does not have.
How many ships transit the Strait of Hormuz week of August 31?
Polymarket · Resolves Sunday, September 6
Our pick: 20-39
Price at 2:40 p.m. ET, Aug 31: Yes 73c – No 28c
Our call: YES, Geek Score 70
This one is arithmetic before it is analysis. Detectable transits have been running at three to five a day through late August. Seven days at that rate gives 21 to 35, which sits inside the 20-39 bin with room on both sides.
The strait has been effectively closed to commercial shipping since February 28, 2026, and the numbers show how complete that is. The EIA put crude and liquids moving through the strait at 4.9 million barrels a day in the second quarter of 2026, against 21.6 million in the fourth quarter of 2025. Pre-crisis traffic ran 85 to 140 vessels a day. War-risk insurance premiums are around 40 times normal, Iranian routing mandates are in force, and a blacklist of 45 vessels is being enforced. Iran and Oman have agreed a revenue-sharing framework, though Tehran says it does not imply reopening.
The risk: the weekend broke the recent calm. US forces struck Iranian rocket launchers preparing to lay mines on Sunday, August 30, a tanker was hit near Oman the same day, and Brent for November pushed above $90 with WTI near $86. Mines actually laid, or a further round of strikes, takes the week under 20 and the call fails. At three points from the market price this is the thinnest call on the board, and the crowd is already close to right.
Summary
| # | Market | Platform | Resolves | Our pick | Our call | Score |
|---|---|---|---|---|---|---|
| 01 | Jobs numbers this month? | Kalshi | Fri, Sep 4 | Above 50,000 | NO | 30 |
| 02 | Unemployment in August | Kalshi | Fri, Sep 4 | Above 4.1% | NO | 35 |
| 03 | Colorado vs. Georgia Tech | Polymarket | Thu, Sep 3 | Colorado (moneyline) | NO | 25 |
| 04 | WILDCHILD: First Week Album Equivalent Units | Kalshi | Sun, Sep 6 | Above 70K | YES | 70 |
| 05 | How many ships transit the Strait of Hormuz week of August 31? | Polymarket | Sun, Sep 6 | 20-39 | YES | 70 |
The widest gap on the board is unemployment, where the crowd pays 60c for a rise to 4.2% and we score it 35, because the falling participation rate that has held the number down since last September has not turned. The two employment calls carry the week; Colorado, WILDCHILD and Hormuz all land within 10 points of the market price, and we have said so in each section rather than dressing up agreement as a forecast.
All prices as of 2:40 p.m. ET, Monday, August 31, 2026. Markets move, and some contracts reset their base intraday, so check the live price before trading.
Prediction markets carry real risk of loss. Nothing here is financial advice; it’s our read on the evidence. Trade responsibly, and if it stops being fun, the National Council on Problem Gambling is at 1-800-GAMBLER.
