MLB Postseason Odds
Who is most likely to win the World Series — and how much does the bracket itself decide? Every playoff team's chance to get through each round, from 20,000 simulations of the bracket.
2026 seasonseries under way — played games count
The bracketas set by MLB
American League
- Wild Card · 6 at 3
- Wild Card · 5 at 4
- Division Series · 1 v 4/51
RaysWinner of 5 at 4 - Division Series · 2 v 3/6Winner of 6 at 3
Odds by round
The Brewers win the World Series in 22% of 20,000 simulations, just ahead of the Dodgers (20%). Even the favourite misses the World Series in 65% of runs; 7 of the 12 teams in the field are under 10% to win it.
Share of 20,000 simulated postseasons in which each team wins the round. Seeds 1 and 2 skip the Wild Card Series.
How this is simulated
This is a run-differential model, not the in-game model. Each team's strength is its Pythagorean win%, pulled toward .500 by 69 games of average play (a full season still carries luck). One game between two teams is Bill James's log5 of their strengths, with a home-field term: between equal teams the home side wins 51.7% — the home win rate in our own play-by-play (model/mlb/win_prob.py).
The formats are the real ones. Wild Card Series: best of three, every game at the higher seed. Division Series: best of five, 2-2-1. League Championship Series: best of seven, 2-3-2, the better seed at home. World Series: best of seven, 2-3-2, the team with the better regular-season record at home. Seeds 1–2 are the two best division winners, 3 the third, 4–6 the wild cards. Series already under way start from their real score.
What it cannot see. Who starts each game — in October a team can lean on its top three starters and its best relievers far more than in July, and nothing here knows that. Nor injuries, nor trade-deadline additions (April runs count the same as September runs). Treat a gap of a few points between two teams as a tie.
Every render uses the same random seed for the season, so the same inputs always print the same numbers.