Kalshi Longshot Picks: Where Our AI Panel Sees 3x The Market
Updated August 10, 2026 · 12 min read · by Jake Hari
If you came to prediction markets from parlay betting, you already know the pull of the cheap ticket. A contract priced at 4 cents pays $1 if it hits, roughly 25 times your stake. What you may not know is which cheap tickets are cheap for a reason and which ones are mispriced. This page is our answer to that question, rebuilt on the same URL as markets settle: every Kalshi market priced under about 20 cents where our price-blind AI panel's estimate lands at a large multiple of the market, with the models' actual stored reasoning and a public grade waiting at the end.
The Quick Answer
Out of every low-priced market our panel has scored in its last five days of runs, seven survived a live re-check. Six are consensus calls, where the whole panel prices the outcome at roughly triple the market: Republicans in the New Hampshire Senate race (11¢ vs a 32.1% blend), Pete Buttigieg for the 2028 Democratic nomination (5.4¢ vs 15.8%), Parker Messick for AL Rookie of the Year (5¢ vs 14.6%), Nolan McLean for NL Rookie of the Year (4¢ vs 14.4%), Péter Magyar for TIME Person of the Year (3¢ vs 9.8%), and Inter to win the Champions League (3¢ vs 8.0%). The seventh is a maverick call: DeepSeek prices Cristopher Sanchez for NL Cy Young at 24.7% against 8 cents, well clear of its own panel-mates. The board, every seat's number, and what each model actually said are below.
Free: The Weekly PM Market Brief — the AI panel's graded record, the week's biggest market-vs-model gaps, and what's spiking next. One email, Sundays. The signup box is at the bottom of this page, or just keep reading: the graded record it promises is printed under the board.
Prices on this page refresh as these markets move, and every figure carries its date. Board updated August 10, 2026; each edition is re-scored in public as its markets settle. Prices as of August 10, 2026; model runs August 5-9, 2026. New to event markets? Start with how sports markets work on Kalshi.
How To Read A Longshot Board
Three things have to be true at once for this page to be worth your time, so let's say them plainly.
First, the payout math. A price in cents is a probability in disguise: a 4-cent contract implies about a 4% chance and pays about 25x when it hits; an 11-cent contract implies 11% and pays about 9x. When our panel says 14% on a 4-cent market, it is not predicting a win. It is saying the ticket is priced like a 1-in-25 shot and reasons like a 1-in-7 shot.
Second, longshots mostly lose. Every market on this board loses more often than it wins, by the market's numbers and by ours. If the panel's blends are exactly right, the expected outcome of this seven-market board is that five or six of the seven settle NO. This is a value board, not a locks board, and nothing on it is a prediction of fact.
Third, cheap contracts are usually bad buys. Decades of racetrack and sportsbook data show bettors systematically overpay for longshots, and the same favorite-longshot bias shows up in event markets. That bias is exactly why this board exists in the form it does: the panel scores every market price-blind, and we only list the rare cheap contract where the models independently break the other way, at 2.5x the market or better. The filter throws away far more than it keeps. This week it threw away over thirty markets and kept seven.
The Board
| Market | Kalshi Price | AI Blend | Top Seat |
|---|---|---|---|
| Republicans win the New Hampshire Senate race | 11¢ | 32.1% | GLM and Sonnet, 37% |
| Pete Buttigieg wins the 2028 Democratic nomination | 5.4¢ | 15.8% | Sonnet, 17% |
| Parker Messick wins AL Rookie of the Year | 5¢ | 14.6% | Three seats at 15% |
| Nolan McLean wins NL Rookie of the Year | 4¢ | 14.4% | Opus, 15.5% |
| Péter Magyar is TIME Person of the Year | 3¢ | 9.8% | DeepSeek, 18% |
| Inter win the 2026-27 Champions League | 3¢ | 8.0% | DeepSeek, 10.3% |
| Cristopher Sanchez wins NL Cy Young (maverick) | 8¢ | 17.4% | DeepSeek, 24.7% |
Prices are the most recent traded prints as of August 10, 2026 (the Inter book last traded August 9); books this thin carry wide spreads, so treat every price as approximate until you see a live quote. Where the market has moved since the panel's run, the current price is shown: Messick ran at 4¢, McLean at 3¢, Magyar and Inter at 2.5¢, Sanchez at 9.5¢, New Hampshire at 12¢, and Buttigieg at 5.45¢, so four of the seven have already ticked toward the models, two have drifted further away, and one is flat.
Every Seat's Number
Panel: Claude Fable, Claude Opus, Claude Sonnet, ChatGPT, Gemini, Kimi, GLM, and DeepSeek, each scoring price-blind from fetched data cards, with a revision round in which every seat read the others' anonymized reasoning. Gemini sat for three of the seven boards; the other four ran with seven of the eight seats, and the blends are equal-weight means of the seats that ran.
| Market | Kalshi | Fable | Opus | Sonnet | ChatGPT | Gemini | Kimi | GLM | DeepSeek | Blend |
|---|---|---|---|---|---|---|---|---|---|---|
| NH Senate: Republican | 11¢ | 27 | 22 | 37 | 36 | 26 | 36 | 37 | 36 | 32.1% |
| Buttigieg 2028 nominee | 5.4¢ | 16.1 | 14.5 | 17 | 16.5 | — | 15.1 | 15.5 | 16 | 15.8% |
| Messick AL ROTY | 5¢ | 14 | 15 | 14.9 | 15 | — | 14 | 14 | 15 | 14.6% |
| McLean NL ROTY | 4¢ | 14 | 15.5 | 14.6 | 15 | — | 14 | 14 | 14 | 14.4% |
| Magyar TIME 2026 | 3¢ | 9.3 | 6.6 | 7.6 | 11.9 | 7.8 | 10.9 | 6.5 | 18 | 9.8% |
| Inter UCL 2026-27 | 3¢ | 7 | 6.5 | 8.5 | 8.6 | — | 8 | 7.3 | 10.3 | 8.0% |
| Sanchez NL Cy Young | 8¢ | 15.8 | 14.7 | 17.1 | 14 | 20 | 15.2 | 18 | 24.7 | 17.4% |
These are model estimates, not predictions of fact and not financial advice, and on the political markets this page takes no side in any race; it analyzes prices. Kalshi is a CFTC-regulated event-contract exchange (18+); you must be in an eligible state to participate, and availability varies by state as of August 2026.
Every seat on this panel is graded against real market settlements — records to date: Claude Fable 82% on 450 graded calls · Claude Opus 84% on 514 graded calls · Claude Sonnet 82% on 498 graded calls · GPT 83% on 3,225 graded calls · Gemini 84% on 3,252 graded calls · Kimi 83% on 2,285 graded calls · GLM 81% on 2,307 graded calls · DeepSeek 80% on 2,332 graded calls. Recomputed daily; the full scoreboard is public.
More live boards from the same panel: New Hampshire Senate 2026 · Republicans 11¢ · 2028 Democratic nominee board · Buttigieg 5.4¢ · TIME Person of the Year 2026 · Magyar 3¢ · Champions League winner odds · Inter 3¢
The Consensus Longshots
These six are the strongest class this board tracks: markets where at least two seats, and here mostly all of them, land at better than double the price. Each entry below quotes or tightly paraphrases the seats' stored reasoning from the run that produced the number.
New Hampshire Senate: The 11-Cent Republican Side
The panel's biggest break from a market anywhere on this board. Kalshi charges 11 cents for Republicans to win the open seat Jeanne Shaheen is vacating; the eight-seat blend from the August 7 run is 32.1%, with five seats at 36-37%. Every seat still calls the Democrats the favorite, which is exactly what makes the price interesting. ChatGPT, at 36%: "The Republican path is still substantial: this is an open seat," with likely nominee John E. Sununu the strongest plausible Republican recruit, a man who has won this seat before. DeepSeek, also 36%, leaned on the polling file: "Pappas leads John E. Sununu 47-44 (UNH June) and 47-41 (Saint Anselm)," margins it read as far too close for a 12-cent side. The panel's own skeptic, Opus at 22%, held the other line: "the president's party lost Senate seats in most postwar midterms," an average of about three per cycle, and that referendum dynamic cuts against the party trying to flip a Democratic-held seat. The full two-sided breakdown lives on our New Hampshire Senate board.
Pete Buttigieg: The Early-State Case At 5 Cents
Kalshi's 2028 Democratic nominee board prices Buttigieg at 5.4 cents in a field with no name above the low twenties. The panel's August 9 run blends him to 15.8%, every seat between 14.5% and 17%. The reasoning is unusually uniform. Sonnet, the top seat at 17%: "Early-state and likely-voter polls plus an explicit intent statement favor Buttigieg over Harris's soft, name-ID-driven national lead." Kimi's version of the same case: early-state likely-voter polls and "the sole explicit intent signal" outweigh national name-recognition leads 27 months before the convention. DeepSeek called it an open race with no favorite above 16%, which is precisely the world where a 5-cent price on a top-tier early-state candidate is the board's mistake. The full 46-name field is on our 2028 Democratic nominee board.
Parker Messick, Cleveland Guardians: 5 Cents For The AL Field's Second Act
The August 9 rookie-board run reads the AL race as Kevin McGonigle's to lose; the seats' shared note calls his lead in the midseason expert poll "growing" and "near-decisive," and Messick's number is what is left over. That leftover is the point: not one of the seven seats that ran this board put the Guardians left-hander below 14%, against a market charging 5 cents. Messick's résumé is doing the arguing: an All-Star first half, the lowest ERA among qualified rookie starters, and a top-three spot with first-place votes in MLB.com's midseason Rookie of the Year poll. A 5-cent price says he needs a McGonigle stumble plus a big September; a 14-15% estimate says that parlay of ordinary baseball events is being sold at a fraction of its value.
Nolan McLean, New York Mets: The Third Man In A Coin-Flip Race
The same run scores the NL race as a true deadlock at the top, the seats' notes citing a literal 144-144 tie between JJ Wetherholt and Sal Stewart in the expert poll. Sonnet's note ends with the phrase that earns McLean his slot here: the NL is a toss-up "with McLean lurking." The Mets right-hander leads all rookies in strikeouts, and ChatGPT's note frames why a pitcher stays live in exactly this situation: modern ROY voters increasingly reward all-around value, "but visible power totals still keep Stewart and Murakami live," which cuts both ways when the two position players split the narrative vote. Every seat holds McLean at 14% or better. The market charges 4 cents, the cheapest price on this board's consensus tier relative to the panel's number: 3.6x.
Péter Magyar: TIME Person Of The Year At 3 Cents
The one non-sports consensus call, and the panel's widest single-seat gap. Magyar's Tisza party won Hungary's April election with 53.6% of the vote, ending Viktor Orbán's 16-year rule, and Magyar has been prime minister since May. Kimi's stored note makes the editorial case in one line: a war year makes Trump the gravitational center, "but repeat-aversion keeps TIME's own anointed Orbán-slayer Magyar nearly level." TIME has picked back-to-back repeat winners almost never, and the panel treats that as structural, not sentimental. DeepSeek went furthest at 18%, six times the 3-cent price, after a revision-round correction described below. Even the panel's low seats, GLM and Opus at about 6.5%, sit at more than double the market. The full 26-outcome board is on our TIME Person of the Year page.
Inter: 3 Cents For A Serie A Champion
Kalshi's Champions League board has no favorite above 16.5 cents, and yet it prices Inter, the reigning Serie A champion, at 3 cents. All seven seats that ran the August 9 club board came in above the market. ChatGPT's note puts Inter's league-phase case alongside the priced elite: Bayern at 89 points, Barcelona at 94, "Inter's 87 points" and a +54 goal difference. Kimi's note lists Inter among "the elite eight" it treats as near-locks for the knockout rounds. Inter won that league by eleven points, have reached two of the last four Champions League finals, and carry the competition's classic quiet-contender profile: nobody's story, everybody's problem. DeepSeek's 10.3% tops the panel; the 8.0% blend is 2.7x the price. Club-by-club numbers are on our Champions League winner board.
The Maverick Call
One market on this board is here on the strength of a single seat, and honesty requires the different framing: one model against the field is a weaker signal than eight models against the market, and this section says so out loud.
Cristopher Sanchez, Philadelphia Phillies: DeepSeek At 3x On NL Cy Young
Kalshi prices Sanchez at 8 cents to win the NL Cy Young in a race Jacob Misiorowski currently owns. In the August 5 run, DeepSeek priced Sanchez at 24.7%, three times the market and nearly five points clear of any other seat. What makes the call worth listing rather than dismissing is that the rest of the panel leans the same direction, just less hard: all eight seats put Sanchez between 14% and 24.7% against the 8-cent price. The mechanism is in the stored notes. Fable, at 15.8%: Sanchez is "the workhorse inheritor," with a board-leading innings pace projecting to roughly 205, a 14-4 record, and a 2.61 ERA, "the profile that wins the moment Misiorowski is capped or fades." Kimi's revision moved him up for the same reason: "the classic workhorse profile that inherits the award in most worlds where Misiorowski is capped, hurt, or regresses." DeepSeek's own revision spent its words trimming the rival fallbacks, cutting Chris Sale on workload risk at the innings floor, and the probability it removed from the rivals landed on Sanchez, up from 22% in the first round. This is a bet on the second scenario of a two-scenario race, priced like a fifth scenario.
The Watch List: What Just Missed
The filter's rejects are as informative as its keeps, so here is what fell out this week and why. Micah Parsons at 3 cents for Defensive Player of the Year drew final seat numbers between 7% and just over 9%, but the panel's notes spend their conviction on Will Anderson Jr. as the breakout challenger, so the number stands without a story. Penn State at 4.5 cents for the Big Ten title opened at 13.7% from DeepSeek, and the revision round walked it back to 9%. Cristiano Ronaldo to Inter Miami at 2 cents drew seats as high as 8.3%, but every stored note argues the likely outcomes are staying at Al-Nassr, retiring, or Sporting CP. And a shelf of 2-cent Rookie of the Year names, TJ Rumfield, Chase DeLauter, and Carson Benge, drew scattered seats at 6-9% with no candidate-specific case. Each of these is one strong revision away from making a future edition; none earned a slot on this one.
Where The Panel Changed Its Mind
Every board here went through a revision round in which each seat read the others' anonymized reasoning, and the arcs are some of the best content the panel produces. The move that most changed this page: DeepSeek's TIME correction. In round one it had loaded 31.5% onto the unlisted field; in round two it wrote that it had "corrected a critical misinterpretation," realizing that a group selection encompassing listed individuals pays the named contracts rather than the field, and redistributed its mass, with Magyar the biggest beneficiary at 18%. Fable's TIME revision was an honest confession of a different kind: "several forecasters priced Netanyahu at 5-10% and I had him off my board entirely; that was an omission, not a judgment." Kimi raised Sanchez from 11.9% to 15.2% on the workhorse-inheritor logic above. And one seat that pointedly did not move: DeepSeek on the Champions League, which held its round-one club distribution and wrote that it was "holding my round-1 distribution essentially unchanged," peer pressure notwithstanding. A model that revises for a stated reason, and a model that refuses to, are both telling you how firm the number is.
What Would Change The Panel's Mind
Every trigger below is a checkable event, and when one fires, this page re-scores.
- New Hampshire: the September 8 primaries confirming or breaking the expected Pappas-Sununu matchup, and the next UNH or Saint Anselm poll. A widening Pappas lead pushes the Republican number down toward the market; another tightening print pushes it up.
- Buttigieg: formal campaign launches by Harris, Newsom, or Buttigieg himself, and the next round of New Hampshire and Michigan early-state polls. His number is built almost entirely on early-state strength; if that fades, so does the 15.8%.
- Messick and McLean: September innings management, any injury to the poll leaders, and the next MLB.com voter poll. Both numbers are second-act bets; a McGonigle or Wetherholt/Stewart stumble is most of the upside.
- Magyar: TIME's shortlist, typically published in early December, is the single decisive trigger. Before that, any autumn news cycle that crowns a new global protagonist cuts his number.
- Inter: the league-phase draw in late August and autumn form. The panel's case is structural, so only results move it.
- Sanchez: any Misiorowski workload news. The entire maverick case is that the favorite's innings are capped or curtailed; a Brewers announcement either way settles the argument early.
Settlement Timeline
| Market | Settles | Next Known Catalyst |
|---|---|---|
| NH Senate | January 2027 (winner sworn in) | Primaries, September 8, 2026 |
| Buttigieg 2028 | November 2028 | Early-state polling waves, fall 2026 |
| AL/NL Rookie of the Year | November 2026 (awards) | September call-ups and innings caps |
| TIME Person of the Year | December 2026 (announcement) | Shortlist, early December |
| Champions League | June 2027 (final) | League-phase draw, late August |
| NL Cy Young | November 2026 (award) | Misiorowski workload decisions |
Graded In Public
This board is an experiment we grade where you can watch. Every number above is stored with its run date, every market's settlement gets scored against the panel's estimate and the market's price, and each new edition of this page opens by re-scoring the last one. When the panel is wrong, that goes on the public scoreboard next to everything else it has ever called. Longshot boards are the harshest possible test of that habit, because most of these markets will settle NO and the question is only whether they settle NO less often than their prices claimed. That is the bet this page makes, in public, on purpose.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $176M traded
- 2028 U.S. Presidential Election winner? · $58M traded
- 2027 Pro Football Champion · $58M traded
- 2028 Republican presidential nominee · $56M traded
- Pro Baseball Champion · $50M traded
- More tech layoffs in 2026 than in 2025? · $31M 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 Bottom Line
Seven markets, all priced 11 cents or cheaper, all still open, all with months of runway, and all carrying model estimates at roughly 2.7x to 3.6x the market, with single seats as high as 6x. If you play them, the honest frame is a portfolio: small, equal stakes across several boards, sized so that the expected five or six losers are survivable and priced-in, never one big swing on the cheapest ticket. The next edition re-scores this one on the same URL.
