What the model sees in Omaha
Eight teams, two double-elimination brackets, one best-of-three final. We simulated the College World Series 10,000 times. The model’s headline: there is no favorite in Omaha this year, and the committee’s last national seed might be the fourth-best team in the field.
Simulations run as of June 8, 2026. These numbers update after every round on the CWS Pulse page as games are played; the figures below are a snapshot.
Every October has a superteam narrative. This June does not. Run the tournament ten thousand times and the most likely champion, North Carolina, wins fewer than one tournament in five. The eighth-most-likely champion, Troy, wins about one in twelve. That 17.4-to-8.3 spread is remarkably flat for an eight-team field: the favorite is barely twice as likely to win it all as the longest shot. If you came here for a lock, the honest answer is that the model doesn’t see one. What it sees instead is structure, and the structure is interesting.
Two brackets, same weight, opposite shapes
Omaha splits its eight teams into two double-elimination brackets of four; the bracket winners meet in a best-of-three final. Sum each bracket’s title odds and they land within a rounding error of each other, 50.2% and 50.0%. The committee, in effect, cut the deck evenly. But the shapes could not be more different.
Bracket 1 is a frontrunner and a fight. North Carolina carries 17.4% title equity and a 32.5% chance of reaching the final, the field’s best on both counts. Behind them: West Virginia (13.4%), Ole Miss (11.1%), and Troy (8.3%). The Tar Heels are the closest thing Omaha has to a favorite, and even they reach the final barely one time in three. Double-elimination is a meat grinder; three of every ten simulated brackets end with UNC going home before the final anyway.
Bracket 2 is a knife fight between near-equals. Georgia (14.8%), Texas (14.4%), and Alabama (11.6%) sit within 3.2 points of each other, with Oklahoma (9.2%) close enough behind to matter. Georgia’s 28.7% and Texas’s 28.6% chances of reaching the final are statistically the same number. Whoever emerges from this side will have earned it twice over, which is exactly why no single team on this side can carry high title odds, however good they are.
The committee seeded West Virginia 16th. The model disagrees.
The most pointed disagreement between the selection committee and the simulation: West Virginia entered the tournament as the #16 national seed, the last protected seed in the field, and the model now gives the Mountaineers the fourth-best title odds in Omaha (13.4%), ahead of the #7 national seed (Alabama) and both 2-seeds. Some of that is the path they realized: WVU won the Morgantown Regional and then won a best-of-three from UCLA. Most of it is the rating itself. The model’s team strength for West Virginia simply sits higher than a 16th seed implies, and Omaha is where that disagreement gets graded in public.
And then there’s Troy, a 3-seed that wasn’t supposed to leave its own regional, which it entered as a guest. The Trojans won the Gainesville Regional on Florida’s field, then took a best-of-three from Southern Miss to reach Omaha. The model gives them 8.3%: the longest odds in the field, and still a number that says this happens about one June in twelve. Three-seeds crashing Omaha is rare. Three-seeds with non-trivial title equity once they arrive is rarer.
How they got here
The eight Super Regionals resolved as: North Carolina over Texas A&M, West Virginia over UCLA, Ole Miss over Auburn (after going through Lincoln in the regional round), Troy over Southern Miss, Georgia over Mississippi State, Texas over Oregon, Alabama over Florida State, and Oklahoma, a 2-seed that first won the Atlanta Regional, over Kansas. Five of the eight national seeds that reached the Supers converted; the field that emerged carries five 1-seeds, two 2-seeds, and Troy.
The most likely final, read straight off the bracket leaders, is North Carolina vs. Georgia. But “most likely” is doing modest work in a tournament this flat: the model’s most probable single final pairing still happens in well under one simulation in ten. The honest summary of Omaha 2026 is a coin flip between bracket halves, a favorite who loses more brackets than he wins, and a 16-seed the model thinks the committee got wrong.
We’ll find out. The numbers above are frozen as of June 8; the CWS Pulse page re-simulates after every round as results come in. When Omaha ends, the forecast’s full history, including this snapshot, stays on the books, graded by what actually happened.
Methodology notes
The engine simulates the entire remaining tournament 10,000 times from the current state. Each simulated game is decided by per-game win probabilities derived from BBMI team ratings; bracket mechanics (double-elimination, the best-of-three final) are modeled explicitly.
Probabilities are frequencies across simulations: “17.4%” means North Carolina won the title in 1,740 of 10,000 simulated tournaments. It is not a prediction that they will win, and it is not a price.
What the model does not know: starting-pitcher matchups and rotation order, day-to-day injuries and availability, and weather. In a short double-elimination format these matter; they are part of why no team’s odds are extreme in either direction.
Reach-the-final percentages sum to roughly 200% across the field (two finalists per simulation); title percentages sum to roughly 100%. Bracket title-equity figures are arithmetic sums of the published per-team odds.
Selection disclosure: this is a forecast published before the games are played. It will look right or wrong in two weeks, and the page it draws from keeps the receipts either way.