Model reveals 2026 March Madness bracket picks
Get SkimNews daily
A curated daily digest across 7 categories: geopolitics, tech, finance, health, energy, sports, culture. Free.
- SportsLine's projection model simulated every game in the 2026 NCAA Tournament 10,000 times and has beaten over 91% of all CBS Sports brackets in four of the past seven tournaments, including correctly picking all four Final Four teams in 2025 and 12 Sweet 16 teams last year.
- Duke earned the No. 1 overall seed in the 2026 bracket, led by freshman Cameron Boozer, who ranks in the top 15 nationally in scoring and rebounding.
- The model picks No. 9 Saint Louis to upset No. 8 Georgia in the Midwest first round; the Billikens finished 28-5 with an 18-game winning streak and were co-champion in the Atlantic 10, with seven players averaging more than nine points per game.
- No. 5 Texas Tech is projected to defeat No. 4 Alabama in the Midwest second round, as Alabama could be without guard Aden Holloway — arrested Monday and under university investigation — and is coming off a loss to No. 15 seed Ole Miss in the SEC Tournament.
- Texas Tech ranks fifth in Division I in 3-point shooting percentage (39.3%) and 3-pointers made per game (11.5), and is in its third consecutive 20-win season under head coach Grant McCasland, the only Red Raiders coach to reach the NCAA Tournament in each of his first three seasons.
- Other stars in the 2026 bracket include freshmen AJ Dybantsa (BYU), Darius Acuff Jr. (Arkansas), and Keaton Wagler (Illinois), plus veterans Braden Smith (Purdue), Zuby Ejiofor (St. John's), and Jaden Bradley (Arizona).
Why it matters: The model's edge — 25 first-round double-digit seed upsets nailed since 2016 and 91%+ bracket accuracy in four of the last seven tournaments — gives data-driven bracket pickers a concrete track record to lean on heading into the 2026 tournament. The Saint Louis and Texas Tech picks point to specific matchup advantages (scoring balance, elite perimeter shooting) that could decide pool-winning brackets, while Alabama's Holloway arrest is a human variable no model can price in.
