March Madness 2027 Odds: The Full Kalshi Championship Board, Scored

The Quick Answer

The March Madness 2027 odds board on Kalshi lists 73 teams for the men's national championship, with co-favorites: Duke and Florida, 12.5 cents each. Our eight-model panel splits that pair hard. It keeps Duke first among named teams at 8.0% while calling the price rich, and it cuts Florida to 4.3%, the deepest fade on the board. Where the panel is above the market: Houston (3.2 cents, blend 7.2%), Arizona (3 cents, 6.2%), and the 48-team "any other listed team" bucket, which the panel prices at 18.4% against a combined 13.4 cents. The full board, every seat's number, and the dark horses the panel refuses to zero are all below.

New to Kalshi event markets? Start here →

This page updates at each scheduled re-score and whenever news moves the board: prices re-stamp with their fetch date, and the tables below are dated August 1, 2026.

The Market At A Glance

Kalshi's 2027 men's college basketball championship market (event KXMARMAD-27) lists 73 programs as separate yes/no contracts. Each one settles on a single rule: if that program is the 2026-27 Division 1 Men's College Basketball National Champion, the contract resolves Yes. The market opened April 7, 2026, the day after Michigan cut down the nets, and settles by April 13, 2027, closing early once a champion is decided. About 5.2 million contracts have traded across the event as of the August 1 fetch.

One structural point worth knowing before the table: Division I basketball runs hundreds of programs deep, and only 73 of them are listed here. If an unlisted program wins it all, every contract on this board settles No. The panel puts real weight on that outcome, and on the 48 listed programs outside the top 25, and both numbers are in the board below.

Two timestamps for everything that follows: panel verdicts were generated the morning of August 1, 2026, price-blind, and every price in this article is as of August 1, 2026, fetched from live Kalshi order books at 11:41 UTC. The Kalshi March Madness board moves daily, so treat the cents columns as a snapshot.

What The Fetched Record Says

Every model received the same fetched data card and no prices. The card carried the 2026 result: Michigan, a No. 1 seed, beat No. 2 seed UConn 69-63 on April 6, 2026 at Lucas Oil Stadium in Indianapolis, in front of 70,720, behind Most Outstanding Player Elliot Cadeau. It was Michigan's first title since 1989 and the first championship ever won with a starting five made up entirely of players who began their careers elsewhere. Dusty May's team also became the first program to score at least 90 points in five games of a single tournament, and the loss was UConn's first in seven trips to the title game.

The card also carried the recent championship line: Kansas in 2022, UConn in 2023 and 2024, Florida in 2025, Michigan in 2026. Two things in that list shaped the panel's reasoning. UConn's back-to-back proves repeats still happen in the portal era. And four different programs won the last five titles, each with a differently built roster, which is why the panel priced coaching machinery over brand names everywhere below.

March Madness 2027 Power Rankings: The Favorites And The Full Board

The headline blend averages all eight seats' revised numbers, after a revision round where each model read the others' reasoning. Gap is the blend minus the market price.

Model estimates generated August 1, 2026, price-blind from live fetched data. Models are frequently wrong; the market price reflects real traders' money.

The remaining 48 listed programs each trade at under 1.2 cents and are grouped in the bucket line; the full list runs from Purdue, Indiana and Villanova at the top down to Richmond, San Francisco and Loyola Chicago at 0.1 cents.

Duke: The Favorite The Panel Keeps, At A Discount

Seven of eight seats put Duke in their top two among named programs in round one, the eighth moved it to second in revision, and the argument was the same everywhere: the machine, not the roster.

"Duke pairs the most reliable elite-recruiting pipeline in the sport with a staff that now retains and supplements with targeted portal veterans, so its floor is a top-10 KenPom team before a ball is tipped." — Claude Opus

The catch is the price. An 8.0% blend against 12.5 cents means the panel thinks the market is paying a brand premium of about a third. Nobody on the panel called Duke a fade; every seat simply found the same probability at a lower number than the crowd's.

Florida: The Co-Favorite The Panel Fades Hardest

Florida shares the top price on the board and finishes eighth in the blend. That 8.2-point gap between 12.5 cents and 4.3% is the single largest disagreement on the board, and it survived the revision round, where Florida's blend barely moved. The panel's logic: the 2025 title was won by a veteran-peak roster whose specific parts are gone, and paying a co-favorite price two years later requires evidence the panel says it does not have. Even the seat friendliest to Florida, which put it at 5.8%, made a case that tops out well below the market:

"They won the 2025 title, have a top-3 NIL collective, and Todd Golden's system translates year-over-year regardless of roster churn — the market will sleep on them because they aren't a blue-blood brand." — DeepSeek V4

When the most bullish seat on the panel prices a team at less than half its market number, that is not a split verdict. That is the board's clearest fade.

Houston: The Board's Biggest Positive Gap Among Names

At 3.2 cents against a 7.2% blend, Houston is where the panel most wants the market to be wrong in its favor. Every single seat priced Houston between 6.0% and 8.0%, the tightest agreement on any contender, and the reasoning never mentioned any individual player.

"Houston is the most system-stable title profile because Kelvin Sampson teams defend, rebound, and avoid empty possessions regardless of roster churn." — ChatGPT (GPT-5.5)

That is the whole thesis: a program whose adjusted-efficiency floor has survived total roster turnover does not need its offseason graded to be priced like a contender.

Arizona: The Revision Round's Discovery

Arizona opened the panel at 5.6% and finished at 6.2%, the largest upward move on any named team in the revision round, against a 3-cent price. The case started with the fetched card, which shows Arizona reached the 2026 Final Four as a No. 1 seed, and got specific from there:

"Their 2026 Final Four run leaned on a freshman core (Koa Peat, Brayden Burries) that, unlike typical champions' rosters, can plausibly return as sophomores under Tommy Lloyd — returning production from a Final Four team is the single strongest preseason predictor." — Claude Fable

One seat doubled its Arizona number after reading that argument, from 3.0% to 6.0%, writing that returning production from a Final Four team "is a concrete edge I had underweighted." The panel's caveat is baked into the wording: this edge depends on those players actually staying, which the card could not confirm.

UConn And Michigan: Process Against Price

UConn is the board's quiet agreement. At 6.6 cents against a 7.5% blend, the panel and the market are close, and the panel's number rests on the deepest track record on the board: per the fetched champions list, Dan Hurley's program won in 2023 and 2024 and returned to the title game in 2026. Several seats raised UConn in revision on the same argument, that a six-point loss in a final is weak evidence against a process that keeps producing title-game rosters.

Michigan, the defending champion, trades at 3 cents, and the panel says 5.5%, well above the market but far below where a title defense might be priced in another era. The structural read:

"A starting five composed entirely of transfers is the least renewable roster in the sport, so Dusty May is re-solving the problem from scratch, and I am pricing the method, not the banner." — Claude Opus

The panel trimmed Michigan in revision on exactly that renewability point. Worth noting against it: the same fetched champions list shows UConn going back-to-back in 2023 and 2024, so repeating is rare, not extinct.

Inside The Revision Round

After round one, each seat read the other seven's anonymized reasoning and could revise with a stated reason. Average cross-model disagreement per board line fell from 0.78 to 0.40 points, and the moves that mattered came with arguments worth quoting.

The biggest was the field bucket. Seats that had priced the 48 unnamed programs cheaply capitulated to bottom-up math:

"I raised Any Other Listed Team from 8% to 16% after multiple forecasters correctly emphasized the depth of Auburn, Purdue, Iowa State and others, and I now believe my original number deeply underweighted that tier." — DeepSeek V4

The round also caught a false premise. One seat's round-one Duke case leaned on a star's return that never happened, and a peer refused to follow it:

"All seven forecasters put Duke top-2 on the Scheyer floor argument I had underweighted, but I stayed below the group's high because H's Flagg-return premise is factually wrong." — Kimi K3

[Editor's note: verified. Cooper Flagg played one season at Duke, declared for the NBA draft in April 2025, and was selected first overall by Dallas in June 2025, per his Wikipedia page as fetched August 1, 2026. The anonymized "Forecaster H" was DeepSeek's round-one case; the blend numbers above reflect the corrected round-two board.]

That is the revision round working as designed: models moved where peers surfaced evidence, and held where the loudest argument was built on a bad fact.

Dark Horses The Panel Won't Dismiss

The whole 73-team field got real numbers, and a few sub-3% lines came with conviction rather than leftover mass:

• St. John's (1.9¢, blend 2.7%): the panel's favorite cheap name. "Rick Pitino ... is a dangerous combination and he's already proven he can build a contender there." — GLM 5.2
• Louisville (3.2¢, blend 2.8%) and Texas Tech (2.7¢, blend 2.4%): priced about right by the market, but the panel's reasoning is worth having. "Pat Kelsey and Grant McCasland are two of the most efficient portal buyers in the country and both have converted mid-tier NIL budgets into top-15 offenses, which is the actual currency now." — Claude Opus
• Alabama (1.7¢, blend 3.7%) and Kentucky (2¢, blend 3.7%): the panel's two quietest edges, both roughly double their price, both argued off SEC roster infrastructure rather than last March's results.
• Any Unlisted Program (blend 1.6%): no contract pays this outcome. The panel holds it on purpose, because a champion from outside the 73 zeroes the entire board.

The Traps: Where The Table Would Mislead You

Transparency beats a clean story, so here is where a raw gap number is not a buy signal.

Texas (5.9¢, blend 1.8%). The board's second-biggest fade after Florida in absolute terms, and proportionally its largest. No seat priced Texas above 2.5%, and the panel's collective view is that the market is paying for momentum its own reasoning could not reconstruct. But two seats separately flagged Texas as the name their round-one numbers most likely undervalued if the offseason broke right, which is exactly the kind of unresolved uncertainty a 1.8% blend hides.

Illinois (7.7¢, blend 3.8%). The market pays for the fetched fact of a 2026 Final Four as a No. 3 seed; the panel discounts it. "Underwood's Euro-import roster model churns hard year to year, so I fade the recency." — Claude Fable. Other seats pushed back and raised Illinois in revision, making this the top of the board's most honest disagreement rather than an edge.

Michigan St. (3.6¢, blend 2.1%) and Arkansas (3.5¢, blend 2.3%). Mid-single-cent prices the panel walks down rather than doubles. No seat zeroed either, and no seat matched the market.

One caveat binds all of these: the panel priced coaching machinery and structure, not August rosters. Several seats flagged their own 2026 offseason knowledge as incomplete, which is precisely why process programs earned the premium and why this board gets re-scored when the season tips.

The 48-Team Bucket, The Longshot Tail, And Why The Bottom Is Noise

The largest single line on the board is not a team. The panel puts 18.4% on the 48 listed programs outside the top 25, against a combined market price of about 13.4 cents. The argument was bottom-up: that bucket contains Purdue, Auburn, Iowa St., BYU, Creighton, Marquette, Wisconsin and Oregon, and pricing each at plausible individual numbers clears the market's total before reaching the tail. One seat put it in historical terms: "champions like 2021 Baylor come from exactly this tier; the public chronically underprices the collective." — Kimi K3

Below that tier, read carefully. Thirty of the 73 contracts trade at 0.15 cents or less on Kalshi's sub-cent tick structure, and at that depth an order-book price is closer to a rounding artifact than an opinion. The per-seat numbers on those rows are leftover field mass, not 73 separate convictions; with eight seats, do not read signal into a 0.1% cell.

How To Read These Prices Eight Months Out

A Worked Example: Duke At 12.5 Cents

A 12.5-cent YES price is not a clean "12.5% chance." Each of the 73 contracts trades in its own order book with a bid-ask spread, so across a mutually exclusive field the listed prices overlap: on this board the 73 mid prices sum to about 108 cents. Normalize Duke's 12.5 against that total and the market's fair-ish implied probability is closer to 11.6%. The panel's 8.0% sits below even the de-vigged number, which is what makes it a real disagreement and not an artifact of the spread.

The same habit applies anywhere a number can be shopped. Bettors comparing a sportsbook future to this board should treat it like shopping a moneyline across books on a live odds screen: the question is never whether a price exists, it is whether a better one does.

Two more honest frictions. First, capital lockup: a YES position here ties up money until the market resolves by April 13, 2027, though unlike a sportsbook future you can sell a Kalshi position any time the book has a bid, which matters if a January roster collapse or a Selection Sunday seed changes your read. Second, thin books: the sub-cent tail trades with wide spreads, and Kalshi charges trading fees, so small edges at the bottom of the board can disappear in execution. If event contracts are new territory, the Kalshi vs DFS pick'em apps breakdown covers the mechanics from a DFS player's frame of reference.

This board reads best next to its siblings: the same eight-seat, price-blind protocol scored the college football 2027 championship board and the NBA 2027 title board, and the cross-sport pattern is consistent: markets pay brands, models pay systems.

Every Seat's Number

The full per-model board from round one, before the revision round moved the numbers above. On a phone this table scrolls sideways; the power-rankings table above carries the headline blend.

Every seat on this panel is graded against real market settlements — records to date: GPT 83% on 1,693 graded calls · Claude Fable 85% on 331 graded calls · Claude Opus 86% on 382 graded calls · Claude Sonnet 84% on 363 graded calls · Gemini 85% on 1,718 graded calls · GLM 80% on 1,874 graded calls · Kimi 84% on 1,863 graded calls · DeepSeek 80% on 1,877 graded calls. Recomputed daily; the full scoreboard is public.

Model estimates generated August 1, 2026, price-blind from live fetched data. Models are frequently wrong; the market price reflects real traders' money.

Scheduled Re-Scores On This URL

An August number on an April market is a starting position, not a verdict. This board gets re-scored on this same page at three known news waves: when the season tips in November and real rosters replace structural priors, after New Year's when conference play has sorted the contenders, and on Selection Sunday, when March Madness odds tighten and the bracket sets the actual paths. After the 2027 champion is crowned, every number above gets publicly graded against the outcome, the same settlement retrospective every board in this series carries.

The Bottom Line

The market and the panel agree the 2027 title chase starts in Durham; they disagree about almost everything else. Kalshi pays Duke and Florida the same 12.5 cents, and the panel splits them by nearly a factor of two. It doubles Houston and Arizona, respects UConn's machine at close to the market's own number, and moves its biggest stack of probability into the 48-team field the market sells for 13.4 cents. This board joins the series next to the fantasy QB1 verdict and the NBA Western Conference 2027 board, and like every board in the series, it is on the public record for grading the morning the nets come down.

More on this: NBA Eastern Conference Odds 2027: LeBron's 76ers Lead At 23\u00a2 · NHL Western Conference Odds 2027: Avalanche Favored At 21\u00a2 · Fantasy RB1 Odds 2026: Gibbs Leads, Bijan Is The Value · DFS Bankroll Rules Applied To Prediction Markets: The Same Discipline, New Surface · The 4-Step Kalshi Prop-Edge Workflow For DFS Players

Hottest Prediction Markets Right Now

• 2028 Democratic presidential nominee · $166M traded
• Rocket Classic Winner · $85M traded
• 2027 Pro Football Champion · $52M traded
• Pro Baseball Champion · $41M traded
• Bitcoin price at the end of 2026 · $30M traded
• 2027 Pro Basketball Champion · $19M traded

Every market above links to our full AI model verdict; browse them all on the Stokastic prediction markets hub.