At 10 a.m. Eastern last Friday, buyers took 21,530 contracts on one of the bleakest questions Kalshi lists: will Major League Baseball play zero regular-season games in 2027? They paid 11 and 12 cents apiece. At 8 a.m. Monday they came back for 16,657 more at 15 cents, and at 8 a.m. Tuesday for another 12,166 at 18 cents. Every one of those trades was a buyer paying the seller's asking price.
Each contract pays $1 if the entire 2027 season is lost and nothing if a single official game is played. Going into Friday, 2,022 of them were open. By 9:13 a.m. Tuesday, 48,712 were.
The fight behind the price has a deadline and a demand. Baseball's labor agreement expires at 11:59 p.m. Eastern on December 1, and the owners have put a salary cap on the table for the first time since 1994. In late July, ESPN's Jeff Passan said what some owners were telling their front offices: "There are owners that believe there won't be a single game next season, so they're telling their GMs to price that into their trade deliberations."
We gave the same question to nine AI models without showing them a price. They agree with the traders that games are in danger. They part ways on whether the whole season goes.
Updated September 15, 2026 · 12 min read · by Jake Hari
How The Prices Moved Since Friday
Kalshi does not post one number for this question. It lists five separate yes-or-no contracts on how many of the 2,430 scheduled games get played, and each pays $1 if its line is cleared:
- No games at all. The entire 2027 regular season is canceled.
- More than 600 games. That is 40 games per team, a quarter of a season.
- More than 1,215 games. Half a season, 81 per team.
- More than 1,930 games. Just under 129 per team. The strike-shortened 1995 season, at 144 games per team, cleared it with 2,017.
- More than 2,425 games. Every game, give or take a rainout. Each undisrupted season since 2017 has finished between 2,429 and 2,431.
Only games completed as official regular-season games count, games postponed and made up later still count, and the reason games are lost does not matter; the full wording is in Kalshi's contract terms.
The repricing hit all five contracts. In the first week of September, the every-game line traded at 61 and 62 cents and the no-season line at 6 to 8 cents. Friday's burst pushed the no-season price into the low teens, and the every-game line traded between 48 and 53 cents that afternoon. Monday's 8 a.m. hour moved everything at once. Along with the 16,657 no-season contracts, 9,652 every-game contracts traded at 50 cents, every one of them a trader paying up to bet against a full season. Through the rest of Monday:
- No games: 15 cents at 8 a.m. and 18 cents by 11 a.m.
- Every game: 50 cents at 8 a.m., 47 cents by 5 p.m.
- More than 1,930: 73 cents at 8 a.m., 65 cents by 6 p.m.
- More than 1,215: 81 cents at 8 a.m., 76 cents by 11 a.m.
Tuesday's 8 a.m. hour brought the third wave. Buyers took 12,166 more no-season contracts at 18 cents, and 19,990 contracts traded at 17 cents on the "600 or fewer" side of the quarter-season line.
The most recent public development in the talks we could find is ESPN's anonymous survey, published September 2, in which 16 of 21 players said they were firmly against discussing a cap.
New to Kalshi event markets? Our Kalshi MLB guide covers how these contracts pay and what they cost to trade. Prices on this page refresh while the labor story is moving, and every quote carries its time stamp. These are market prices and model estimates, not predictions of fact and not financial advice.
Stay ahead of the markets.
Daily insights and expert picks on Kalshi, Polymarket, and what's moving markets.
Free forever. Unsubscribe anytime.
Why This Fight Is Different: The Salary Cap
The last time baseball stopped, the 2021-22 lockout, ended with a higher minimum salary, a $50 million bonus pool for young players, a universal designated hitter and a 12-team postseason. The owners are asking for something bigger this time: a different system.
MLB's proposal, tabled in June, caps team payrolls at $245.3 million starting in 2027, with a floor of $171.2 million. It also limits free-agent deals to five years and $202 million for a player changing teams, and six years and $265 million for one staying put. The union's reply: "Player share under their proposal would go down... players would lose over half a billion dollars."
Both sides have spent the summer saying where they stand. Bruce Meyer, the players' association's interim executive director, calls a cap "a form of institutionalized collusion." Commissioner Rob Manfred answered in July: "I have an ownership group that is more united than any group in my entire time of baseball." A day later, Meyer told reporters, "Salary cap is the ultimate excuse not to compete." President Donald Trump sided with the owners on June 5, per CBS Sports: "If you don't have a salary cap, you don't have a sport."
The union's counterproposal, per the Associated Press, raises minimum salaries, expands revenue sharing and charges a "competitive integrity tax" to teams that do not meet a minimum payroll. It has no cap. On August 27, MLB offered roster changes, including "quiet periods" that limit free-agent signings, but only if the players accept a cap. Meyer called the signing window "a game of musical chairs pitting player vs. player."
Meyer said in March that he would "be shocked if they didn't do a lockout, when the agreement expires," per Yahoo Sports. A lockout on December 2 would freeze free agency and trades, as it did in 2021, and cost no games until spring. ESPN's labor explainer puts that point at early to mid-March. Opening Day is scheduled for Thursday, March 25, 2027. Kalshi's separate contract on a new deal landing before December 2 trades at 5 cents bid, 7 cents ask.
What Happened The Last Times Owners Wanted A Cap
Labor fights have cost North American leagues games before, and the cause of each one matters more than its length. MLB season totals below are from MLB's own records.
| Stoppage | What It Was About | Length | What It Cost The Season |
|---|---|---|---|
| MLB 1990 lockout | Owners' salary cap proposal | 32 days | Nothing: the full schedule was kept (2,105 games) |
| MLB 1994-95 strike | Owners' salary cap proposal | 232 days | 1994 stopped at 1,600 games; 1995 ran 144 per team (2,017 games) |
| MLB 2021-22 lockout | Minimum pay, young players' pay | 99 days | Nothing: canceled games were made up, all 2,430 played |
| NHL 2004-05 lockout | Owners' salary cap demand | 310 days | The whole season: 0 of 1,230 games |
| NBA 1998-99 lockout | Revenue split, cap structure | 204 days | 50 of 82 games per team |
| NBA 2011 lockout | Revenue split, a hard cap | 161 days | 66 of 82 games per team |
| NHL 2012-13 lockout | Revenue split inside an existing cap | 119 days | 48 of 82 games per team |
Twice before, MLB owners put a cap on the table. In 1990 the lockout lasted 32 days and ended without one. Opening Day moved back a week, and the season ran three days long to fit every game. In 1994 the players struck on August 12, the rest of the season and the World Series were canceled on September 14, and the strike ran 232 days until Judge Sonia Sotomayor's injunction on March 31, 1995. It also ended without a cap.
The one North American league ever to lose a whole season, the NHL in 2004-05, lost it over a cap. The owners demanded what they called "cost certainty," the players rejected six league proposals as caps by another name, and the season was canceled on February 16, 2005. The players came back to a $39 million cap and a share of league revenue cut from 76% to 54%. The NHL's lost season is the whole case for the no-season contract.
The offseason baseball lockouts point the other way for the every-game contract. Both 1990 and 2021-22 ate into spring, and in 2022 MLB had already canceled the first two series on March 1 before the March 10 deal. Those games were made up. A late-March settlement with a rebuilt schedule would still pay the every-game line, even with a late Opening Day.
The Board
We handed the question to nine AI models: Claude Fable, Claude Opus, Claude Sonnet, ChatGPT, Gemini, Grok, Kimi, GLM and DeepSeek. None saw a price. Each model split 100% across six ranges of games played, from none at all to every game, and all five Kalshi lines were read off that one forecast. Then each read the others' reasoning with the names hidden and chose whether to revise. The table shows the panel's middle answer beside Kalshi's live quote and the range each line traded in the first week of September.
| Line | Kalshi midpoint now | Panel | Bid / ask now | Traded Sept. 1-7 |
|---|---|---|---|---|
| No games at all | 19.5¢ | 6% | 19 / 20 | 6 to 8 |
| More than 600 | 81.5¢ | 87% | 81 / 82 | 91 to 92 |
| More than 1,215 | 75.5¢ | 76% | 75 / 76 | 81 to 83 |
| More than 1,930 | 65.5¢ | 60% | 65 / 66 | 74 to 76 |
| Every game (more than 2,425) | 47.5¢ | 43% | 47 / 48 | 61 to 62 |
Kalshi quotes as of September 15, 2026, 9:13 a.m. ET (Tuesday). "Traded Sept. 1-7" is the range of daily closing prices in the first week of September, in cents. The panel figure is the panel blend: the seat median of the nine models after the revision round. These are market prices and model estimates, not predictions of fact and not financial advice.
Start with the top row, because it is where the panel and the traders part company. The no-season contract sits at 19.5 cents. The panel's middle answer is 6%, and not one of the nine seats went above 12%.
On the rest of the board the panel runs close to the market, and a shade darker on a full season. Its every-game number is 43% against a 47.5-cent midpoint, and its 1,930 line is 60% against 65.5 cents. The models expect games to be lost. They do not expect the season to be lost.
Read the traders' five prices together and they make a specific bet: either there is no 2027 season, or there are more than 600 games. The no-season and more-than-600 midpoints add up to 101 cents, one cent over a dollar, which is the width of the gap between what buyers bid and what sellers ask. Read at the midpoints, that leaves nothing for a season of 1 to 600 games. The panel leaves 7% there and puts its median season at 2,200 games, which works out to 146.7 per team.
Every number in this piece gets graded in public once the season settles, on the full graded scoreboard.
Every Seat's Number
The per-model numbers after the revision round show where the nine actually disagree: how long a stoppage runs once it starts.
| Model | No games | More than 600 | More than 1,215 | More than 1,930 | Every game | Median games |
|---|---|---|---|---|---|---|
| Fable | 8% | 89% | 80% | 67% | 47% | 2,390 |
| Opus | 10% | 86% | 76% | 60% | 38% | 2,200 |
| Sonnet | 5% | 89% | 79% | 64% | 43% | 2,265 |
| ChatGPT | 3% | 94.5% | 89% | 75% | 51% | 2,428 |
| Gemini | 6% | 83% | 70% | 54% | 45% | 2,100 |
| Grok | 11% | 84% | 75% | 58% | 35% | 2,110 |
| Kimi | 5% | 92% | 84% | 72% | 55% | 2,430 |
| GLM | 5% | 87% | 74% | 56% | 33% | 2,000 |
| DeepSeek | 12% | 80% | 62% | 35% | 10% | 1,533 |
| Panel Median | 6% | 87% | 76% | 60% | 43% | 2,200 |
| Kalshi Midpoint | 19.5¢ | 81.5¢ | 75.5¢ | 65.5¢ | 47.5¢ |
Every seat on this panel is graded against real market settlements — records to date: Claude Fable 87% on 4,062 graded calls · Claude Opus 86% on 4,151 graded calls · Claude Sonnet 85% on 4,127 graded calls · GPT 85% on 10,808 graded calls · Gemini 86% on 10,117 graded calls · Grok 84% on 8,483 graded calls · Kimi 85% on 5,318 graded calls · GLM 83% on 4,898 graded calls · DeepSeek 83% on 4,787 graded calls. Recomputed daily; the full scoreboard is public.
Every seat put a lockout in effect on December 2 at 88% or higher, which fits the 5-to-7-cent price on the separate contract that pays only if a new deal arrives first. The split comes after it. DeepSeek alone puts its median season at 1,533 games and gives every game 10%.
At the other end, ChatGPT and Kimi put their median seasons at 2,428 and 2,430 games, a full schedule, and give every game 51% and 55%. Both lean on 1990 and 2022, when spring lockouts ended in time to play every game.
More live boards from the same panel: Mike Trout next team odds, where the Angels are 84¢ bid, 86¢ ask · World Series 2026 odds, where the Dodgers are 30¢ bid, 30.3¢ ask · AL pennant odds 2026, where the Yankees are 25¢ bid, 27¢ ask. Prices fetched September 15, 2026.
Where The Panel Changed Its Mind
In the revision round each model read the others' reasoning, labeled only Forecaster A, B, C and so on. Eight of the nine revised, all by small amounts. The panel's no-season median went from 8% to 6%, and its every-game median from 45% to 43%.
Gemini cut its no-season number from 9% to 6% after reading two forecasters who pointed back to the court order that ended the 1994-95 strike:
"convinced by Forecasters D and F that an NHL-style total wipeout underweights the 1995 Sotomayor injunction precedent." — Gemini
ChatGPT went the other way, from 1% to 3%, and the reason was the owners' own words as Passan reported them:
"Forecaster H had the strongest opposing argument: Passan's report that some owners are already pricing zero games makes wipeout more than a token tail." — ChatGPT
Fable trimmed its every-game number from 49% to 47% over a detail of the 2022 ending. Opening Day slipped a week that year and every game was still made up, but a deal that lands later than that may not get the same treatment:
"Forecasters B and E showed a deal landing late March to early April only gets games made up perhaps half to three-quarters of the time, whereas I treated any pre-April deal as full restoration." — Fable
DeepSeek held every number, including 10% on every game, the lowest on the panel:
"With the union calling a cap “institutionalized collusion,” a quick fold is implausible; the NHL 2004-05 template better fits a genuine first-cap fight than 1990’s superficial retreat." — DeepSeek
[Editor's note: MLB owners also proposed a salary cap in 1990 and 1994; the 2026 proposal is the first since 1994.]
What Would Change The Panel's Mind
Three dates and decisions, each one checkable, each with the direction it pushes.
- December 1 At 11:59 p.m. Eastern. A deal before the agreement expires would lift every line toward a full season. A lockout is already in the numbers (every seat puts it between 88% and 95%), so what moves the board that week is whether the talks keep going after the doors close.
- Any Movement On The Cap. The owners trading the hard cap for a tax-and-floor system, or the union agreeing to bargain over any limit on payrolls, removes the cause behind 1994 and the NHL's lost season. Pushes the every-game line up and the no-season line down.
- Mid-March 2027. If MLB cancels regular-season games with no deal in hand, the question becomes whether they get made up, as in 2022. Cancellations without a deal push the every-game line down hard and bring the 1,930 and 1,215 lines into play.
Settlement Timeline
The contract settles on MLB's official count of completed regular-season games. The dates below are the ones that move it.
| Labor Deal Expires | 11:59 p.m. ET, Tuesday, December 1, 2026 |
| Games Start To Be Lost | Early to mid-March 2027, per ESPN's labor explainer |
| Scheduled Opening Day | Thursday, March 25, 2027 |
| Scheduled Final Day | Sunday, September 26, 2027 |
| Settles | Expected Sunday, October 31, 2027, and no later than November 7 |
This page is re-scored when the story moves. Prices last read September 15, 2026, 9:13 a.m. ET.
More on this: Kalshi World Series Odds 2026: The Dodgers Surged. The Panel Held Firm. · MLB Playoffs 2026 Odds: Who Makes It And Who Wins It All · MLB Worst Record Odds: The Race Nobody Wants To Win
The Bottom Line
Baseball's labor deal runs out on December 1, the owners want a salary cap for the first time since 1994, and the union will not discuss one. Since Friday morning, traders have bought the bleakest outcome, a 2027 with no games at all, in three waves at prices that climbed from 11 cents to 18. Open contracts on it went from 2,022 to 48,712.
Nine models that never saw those prices agree that games are in danger, and they put a full season a little below the market, at 43%. They part ways on the wipeout: 6% from the panel, 19.5 cents from the traders.
December 1 is the next scheduled deadline. Mid-March is when games start to be lost, and when the lower lines come into play.
Hero photo: a Billy Loes Model A2121 Wilson baseball glove, public domain, via the Digital Public Library of America and Wikimedia Commons.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $218M traded
- 2027 Pro Football Champion · $110M traded
- Fed decision in Sep 2026? · $78M traded
- Pro Baseball Champion · $73M traded
- 2028 U.S. Presidential Election winner? · $66M traded
- 2028 Republican presidential nominee · $64M 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.
