Updated September 13, 2026 · 12 min read · by Jake Hari
A 3-cent contract pays $1 if it hits. That payout is why the cheap end of Kalshi pulls in so many people who came over from parlays. The catch is that cheap contracts usually lose, and our own grading says an AI model disagreeing with the market is a weaker signal than it looks. This page holds both facts at once. Every few weeks we log the markets priced at a nickel or less where six or seven AI models, each pricing the race without being shown the market, come back at several times the price. Then we show exactly what the models said, and track whether each disagreement holds up, starting with how the last batch aged.
The Quick Answer
Four new longshots make this edition, each quoted at 5 cents or less: Colorado State and Fresno State to win the Pac-12, and Zac Taylor and Andy Reid for NFL Coach of the Year. On all four, the panel's middle number lands at two and a half to four times the asking price. None of the last edition's seven picks has settled; two have moved toward the panel, three have moved away, and two have barely budged. The board, every seat's number, and the models' own words are below.
Prices on this page refresh as these markets move, and every figure carries its date. These prices were fetched at 3:54 AM Central on September 13, 2026, the same morning the panel ran. New to Kalshi event markets? Start here →
Stay ahead of the markets.
Daily insights and expert picks on Kalshi, Polymarket, and what's moving markets.
Free forever. Unsubscribe anytime.
How The Last Edition Aged
The August 10 edition listed seven markets. Nothing on it has settled, so there is no win-loss record to print yet. What we can grade is direction: if the models were reading something the market had missed, prices should drift toward them as news arrives.
| Last Edition's Pick | Price, Aug. 10 | Panel, Aug. 10 | Now (Bid / Ask) | Direction |
|---|---|---|---|---|
| Republicans win the New Hampshire Senate race | 11¢ | 32.1% | 15¢ / 16¢ | Toward the panel |
| Parker Messick, AL Rookie of the Year | 5¢ | 14.6% | 12¢ / 13¢ | Toward the panel |
| Pete Buttigieg, 2028 Democratic nominee | 5.4¢ | 15.8% | 5.4¢ / 5.5¢ | Flat |
| Inter, 2026-27 Champions League | 3¢ | 8.0% | 2¢ / 3¢ | Flat |
| Nolan McLean, NL Rookie of the Year | 4¢ | 14.4% | 1¢ / 2¢ | Away |
| Péter Magyar, TIME Person of the Year | 3¢ | 9.8% | 1¢ / 2¢ | Away |
| Cristopher Sanchez, NL Cy Young | 8¢ | 17.4% | 1¢ / 2¢ | Away |
The August panel figures are shown as published then, an equal-weight average of the seats; this edition uses the middle seat instead.
Parker Messick is the clean case for the models. His AL Rookie of the Year contract has more than doubled since August, to a 12-cent bid. That is most of the way to where the panel sat.
The New Hampshire Republican side rose too, to a 15-cent bid, with a twist worth knowing. When the panel re-scored that race on August 23, its own middle number fell to about 13.5%, a hair below where the market now trades, and our New Hampshire Senate board carries that full two-sided read. Price and panel both moved, and they met in the middle.
The three that went the wrong way now share one price: a 1-cent bid and a 2-cent ask. Sanchez fell the furthest, from 8 cents. McLean is the one to watch, because the panel's most recent run on the NL race, on August 23, still had him at about 18.5% while the market walked the other way the entire time. Our NL Rookie of the Year board has the whole field.
Three wrong-way moves out of seven is roughly what our own data predicts. When we graded 1,861 settled contracts priced between July 20 and August 10, the market beat the panel. On the 147 contracts where the two disagreed by 15 points or more, the panel's side won 27 times, or 18.4%. The same study of where prediction markets are sharp found that contracts under 15 cents in thin books ran about 4 points too rich, so cheap tickets were, on average, priced above their real hit rate. Read everything below as a disagreement with a documented, mostly losing record behind it.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $216M traded
- 2027 Pro Football Champion · $97M traded
- Pro Baseball Champion · $71M traded
- 2028 U.S. Presidential Election winner? · $66M traded
- 2028 Republican presidential nominee · $64M traded
- Fed decision in Sep 2026? · $62M traded
Every market above links to our full AI model verdict; browse them all on the OddsShopper prediction markets hub, and see how every settled call actually scored on the full graded scoreboard.
The Board
Seven of the panel's eight seats priced the Pac-12 race and six priced the Coach of the Year race. Each worked from records, schedules, and settlement rules fetched the morning of September 13, and none was shown a market price. Each then read the others' anonymized reasoning and could revise. The number below is the middle seat after that revision round.
| Longshot | Kalshi (Bid / Ask) | Panel Median | Seat Range |
|---|---|---|---|
| Colorado State wins the Pac-12 | 3¢ / 4¢ | 12.0% | 11% to 14.5% |
| Fresno State wins the Pac-12 | 4¢ / 5¢ | 12.5% | 12.5% to 13% |
| Zac Taylor, NFL Coach of the Year | 2¢ / 3¢ | 11.5% | 10.8% to 13% |
| Andy Reid, NFL Coach of the Year | 1¢ / 3¢ | 9.2% | 7.5% to 9.5% |
Read the Fresno row first, because it is the strangest. Seven models landed within half a point of one another on a team the market prices at a nickel or less.
Every one of these books is thin, and the ask is what a buyer actually pays, so the multiples in this piece are measured against it. That works out to about 3x for Colorado State, 2.5x for Fresno State, 3.8x for Taylor and 3.1x for Reid. Reid's spread is the widest: a buyer at 3 cents could only sell back at 1, with two-thirds of the stake gone to the gap.
Every Seat's Number
| Longshot | Ask | Fable | Opus | Sonnet | ChatGPT | Gemini | Kimi | Grok | Median |
|---|---|---|---|---|---|---|---|---|---|
| Colorado State, Pac-12 | 4¢ | 11 | 12 | 13 | 13.5 | 14.5 | 11.5 | 12 | 12.0% |
| Fresno State, Pac-12 | 5¢ | 12.5 | 13 | 12.5 | 12.5 | 12.5 | 12.5 | 12.5 | 12.5% |
| Zac Taylor, Coach of the Year | 3¢ | 11.5 | 11.5 | 13 | 10.8 | 11 | 12 | — | 11.5% |
| Andy Reid, Coach of the Year | 3¢ | 9.5 | 9 | 9.5 | 9.3 | 7.5 | 9 | — | 9.2% |
Every seat on this panel is graded against real market settlements — records to date: Claude Fable 88% on 3,267 graded calls · Claude Opus 88% on 3,356 graded calls · Claude Sonnet 86% on 3,332 graded calls · GPT 86% on 9,999 graded calls · Gemini 87% on 9,161 graded calls · Kimi 86% on 4,636 graded calls · Grok 84% on 7,594 graded calls. Recomputed daily; the full scoreboard is public.
Every number in this piece gets graded in public once its market settles, on the full graded scoreboard.
More live boards from the same panel: Pac-12 Championship odds · Boise State 46¢ bid · 2028 Democratic nominee board · Buttigieg 5.4¢ bid · Champions League winner odds · Inter 2¢ bid · Prices fetched September 13, 2026
Dark Horses That Just Missed
Two names drew real numbers and still fell short of the bar. Washington State's middle number dropped to 5% after an 0-2 start, under 2x its 3-cent ask. Dan Campbell sits at 6.5% for Coach of the Year against a 4-cent ask, under 2x, though Opus kept him on its short list for a concrete reason: "Campbell (easiest schedule, 9-8 baseline) and Schottenheimer are the next tier." Detroit's schedule is the lightest in the league by the exchange's own win totals for each opponent.
Colorado State And Fresno State: Two Ways Into A Brand-New Pac-12
This Pac-12 is a new league in an old name. Six of its eight programs are in their first season in it, and the other two, Oregon State and Washington State, spent 2024 and 2025 without a conference. Everyone plays a seven-game league round robin, the top two meet on Friday, December 4, and the higher seed hosts. Every seat agrees Boise State is the team to beat: the Broncos took 14 first-place votes in July's media poll, then led No. 2 Oregon at halftime before losing 34-27 as 24.5-point underdogs. Kalshi has Boise at a 46-cent bid; the panel's middle seat says 43%.
The disagreement is underneath Boise, in a middle tier the market has spread thin.
Colorado State Rams: 3 Cents After A 2-0 Start
Colorado State went 2-10 last season and was picked to tie for sixth in July. Under first-year coach Jim Mora, it has since beaten Wyoming 35-13 and FCS Southern Utah 58-24, and it is 2-0 for the first time since 2011, per the AP game story. Quarterback Hauss Hejny threw five touchdown passes in the opener and accounted for five more touchdowns against Southern Utah.
Gemini made the most of it in its revision round:
"The round-robin format and home-field title game amplify Boise State's clear talent edge, while Colorado State's immediate offensive explosion under Jim Mora upends the preseason poll's middle tier." — Gemini
The honest counterweight is the schedule so far: one Mountain West rival and one FCS opponent. The Rams play BYU on September 19, open league play against Oregon State on October 3, and host Boise State on November 7. That last date is the whole case in one afternoon.
Fresno State Bulldogs: The Panel's Most Unanimous Number
Fresno State was picked third in July, went 9-4 in 2025, and is 1-1 under Matt Entz. The loss was a 39-0 shutout at USC; the win was 49-3 over Sacramento State. The market has the Bulldogs at a 4-cent bid, below Texas State, a team that has lost both of its games.
The panel did not flinch at the USC score. Grok filed it under expected results: "Fresno's 39-0 at USC and Texas State's 59-7 at Texas mostly confirmed mismatches." Before the revision round, Fresno's seven numbers ran from 11.5% to 13.5%; after it, all seven sat at 12.5% or 13%. The Bulldogs host Boise State on October 11 and Colorado State on November 22, so the two home dates that decide their season also settle this row.
Zac Taylor And Andy Reid: Two 6-11 Teams, Two Cheap Coaches
AP Coach of the Year is a win-jump award with a floor. Of the 36 winners since 1990 with a prior season to compare, 32 improved their team by four or more wins, the median jump was six, and the median winning record was 12 wins. Thirteen of the 36 were in their first year with the team, which is why the market's favorite is Jesse Minter, the Ravens' new coach, at a 14-cent bid. The panel agrees he belongs near the top. It parts ways with the price on two coaches whose teams both went 6-11 last year.
Zac Taylor, Cincinnati Bengals: The Biggest Jump The Market Expects
Cincinnati went 6-11 and gave up 492 points. The exchange's separate season-wins market has the Bengals at a median of 10.93 wins, the biggest expected improvement of any team, with Joe Burrow atop the depth chart. By the same season-wins market's numbers for each opponent, the Bengals' schedule is the second-lightest in the league, behind only Detroit.
All six seats put Taylor between 10.8% and 13%, against a 3-cent ask. ChatGPT laid out the whole argument in one line: "Taylor leads because Cincinnati combines a 6-11 baseline, Burrow, the largest expected improvement, light schedule, and clean injury report." Taylor is 57-65-1 across seven seasons in Cincinnati, and the Bengals open at home against Tampa Bay on Sunday, September 13, at noon Central.
Andy Reid, Kansas City Chiefs: The Unluckiest 6-11 In Football
Kansas City outscored its opponents 362-328 last season and still finished 6-11. By points scored and allowed, that roster played like a 9.5-win team, the largest gap between record and scoring anywhere in the league. The season-wins market expects 10.52 wins this year, with Patrick Mahomes at quarterback. Reid won this award once, in 2002 with Philadelphia.
The seats like the rebound and doubt the credit. Opus explained why it kept Reid below the other two:
"I cut Reid below Minter and Taylor because voters may read a Chiefs rebound as a return to normal and credit Mahomes." — Opus
Every seat still landed between 7.5% and 9.5%, against a 1-cent bid. The Chiefs host Denver on Monday, September 14, at 7:15 PM Central.
Where The Panel Changed Its Mind
The revision round, where each seat reads the others' anonymized reasoning, moved the Pac-12 board more than the coaching board. On Colorado State it pulled from both ends. Gemini came down from 18.5% to 14.5%, while Grok, the panel's lowest seat on the Rams at 7%, came up to 12%. Gemini's revised note leaned on a count the card supports, "Hauss Hejny accounting for ten total touchdowns," across the two games.
The best single arc came from Kimi, on the favorite, after another forecaster pointed out a schedule detail it had missed:
"Boise plays four of its seven league games at home, including the likely de-facto semifinal against San Diego State on Nov. 22, and as probable top seed it hosts the title game. That pushes me from ~41 to 44." — Kimi
On Coach of the Year, the models mostly held. Opus wrote that "the other forecasters confirmed my structure rather than changing it." ChatGPT moved a little toward the Giants and Cowboys coaches, and Fable trimmed Minter back, writing that "his 8-9 baseline caps the delta." Taylor and Reid barely moved in either direction, which tells you how firm those two numbers are.
What Would Change The Panel's Mind
Every trigger below is a checkable event, and when one fires, this page re-scores.
- Colorado State: the November 7 home game against Boise State. A win there pushes the Rams' number up sharply; an October loss to Oregon State or Texas State would pull it toward the market well before then.
- Fresno State: the October 11 home game against Boise State. In a seven-game round robin, a Fresno win over the favorite is the single result that moves this row the most.
- Zac Taylor: any injury designation for Burrow, and the Bengals' record against that 10.93-win line through October. A quarterback absence pushes his number down faster than anything else.
- Andy Reid: the Chiefs' first month against their 10.52-win line. The voter-credit worry does not go away, so even a hot start raises his number less than it would raise Taylor's.
Settlement Timeline
| Market | Settles | Next Known Catalyst |
|---|---|---|
| Pac-12 Championship | Title game Friday, December 4, 2026 (Kalshi expected expiration December 6) | League play opens the weekend of October 3 |
| NFL Coach of the Year | AP announcement (Kalshi expected expiration March 30, 2027) | Bengals vs. Buccaneers, September 13; Chiefs vs. Broncos, September 14 |
| MLB rookie and Cy Young awards | Postseason award announcements (Kalshi expected expiration December 31) | Final month of the regular season |
| New Hampshire Senate | Kalshi expected expiration January 4, 2027 | The November general election |
The board is re-scored when the story moves; as of September 13, 2026.
The panel's calls are on the record. Our 8-model AI panel puts a graded, public verdict on the full Coach of the Year field — see the live model verdict, every coach priced by the same seats.
The Bottom Line
Four cheap contracts, and four places where six or seven models reading the same fetched records landed at two and a half to four times the asking price. The record says most disagreements like these go the market's way, and the last edition's early scorecard, two toward and three away, looks a lot like that record. Colorado State, Fresno State, Taylor and Reid are the models' side of the argument, stated in public with the date on it. The next edition opens by re-scoring this one on the same URL, after the Pac-12's first league weekend and a month of NFL football.
Trade prediction markets on Polymarket too.Most readers here already have Kalshi — Polymarket is the other major venue, and code OS4 adds a $50 trading bonus on a $10 deposit: Get the Polymarket bonus →
Affiliate disclosure: OddsShopper earns a commission on signups through this link. Terms as stated by Polymarket; 18+, availability varies by state.
More on this: The Favorite-Longshot Bias Is Backwards On Kalshi Weather — And The Tick Size Explains It · England Vs Pakistan Test Odds: Is The Kalshi Draw Cheap? · Florida Governor Odds: Donalds Vs Jolly After The Primary · 2026 Governor Races: Every Kalshi Board We Track · GTA 6 PC Release Date: The Pattern Rockstar Never Breaks
