🏀 UConn vs. Michigan AI Prediction: Claude's National Championship Picks & Projected Stats

We've already covered the national championship game tonight between the No. 2 seed UConn Huskies and No. 1 seed Michigan Wolverines from every angle, but what do the robots predict will happen tonight?

I trained Claude, the highly advanced AI model from Anthropic, to make its UConn vs. Michigan AI prediction ahead of their clash tonight at 8:50 p.m. ET (TNT/TBS/truTV) - including its exact score prediction, full projected stats and box score, and best player props based on its internal simulation for the title game.

As always, our AI predictions should be viewed as primarily entertainment (with a dash of logic) to complement the rest of our March Madness predictions. You can see our full methodology for how we trained our AI model below. Take it away, Claude!

🤖 UConn vs. Michigan AI picks

All picks and analysis below courtesy of our Claude AI model. See all of our experts' college basketball picks based on the latest college basketball odds.

Michigan wins this game, but it's not the coronation the UConn vs. Michigan odds imply. The Wolverines are historically dominant by KenPom — first in adjusted defensive efficiency, fourth in adjusted offensive efficiency, a +39.72 net rating that ranks second all-time — but they're running into the single most battle-tested program in modern college basketball. Dan Hurley is 11-0 in the Sweet 16 or later. UConn is 6-0 all-time in national championship games. And Tarris Reed Jr. is playing the game of his life against his former team.

I'm projecting a 143-point total in a game where UConn dictates tempo around 65 possessions. Yaxel Lendeborg plays but is limited to roughly 24 minutes with reduced lateral mobility, which narrows Michigan's offensive ceiling. Aday Mara fills the void as the Wolverines' primary scorer. UConn keeps this inside the number — the Huskies cover +6.5 — but Michigan's depth and shooting from five positions is ultimately too much for UConn to overcome.

AI score prediction: Michigan 74, UConn 69

📊 UConn vs. Michigan stats & projected box score

See our lead expert's UConn vs. Michigan prediction for the national championship.

Scoring summary

Michigan's depth advantage is the difference, but UConn controls the terms of engagement. The Huskies drag this game into the mid-60s in possessions — well below Michigan's preferred tempo — and their defensive connectivity limits the Wolverines to 48.2% shooting, the lowest mark Michigan would post in the entire tournament.

UConn keeps this close because Reed dominates the paint and Karaban spaces the floor, but Michigan's ability to score from all five positions eventually creates enough cracks. The Wolverines win the free-throw battle decisively, and UConn's inability to convert at the stripe — a season-long issue for a team ranked 306th in free-throw rate — is the margin between a cover and an outright upset.

Michigan projected stats

Shooting splits: 48.2% FG, 33.3% 3PT, 72.2% FT

Mara is the story on Michigan's side. With Lendeborg limited to 24 minutes of spot-up shooting — the "really good 38-year-old at the YMCA" that Dusty May described — the 7-foot-3 center steps into the alpha role he previewed when he scored a career-high 26 against Arizona. His 16-point, 8-rebound, 3-block line exploits UConn's biggest vulnerability: free-throw rate allowed, where the Huskies rank 307th nationally. Mara gets to the line seven times and alters at least five additional shots at the rim.

Cadeau is the engine with 7 assists fueling Michigan's half-court offense, but the Wolverines' turnovers against UConn's elite point-of-attack defense keep the Huskies in the game all night. Lendeborg's 12-point line reflects his limitations — two spot-up threes, five boards, no drives to the rim, no defensive slides in isolation. His 24 minutes are the fewest he's played in a game he's finished this season. It's enough to keep UConn honest, but this is Mara and Cadeau's game.

UConn projected stats

Shooting splits: 46.3% FG, 31.6% 3PT, 65.0% FT

Reed is the best player on the floor tonight. His 18-point, 12-rebound projection continues a tournament in which he's averaged 20.8 points and 13.0 rebounds on 58.2% shooting — now against his former team, in his final college game. He bullies Mara on the block, draws fouls on Johnson, and wills UConn back into the game every time Michigan tries to pull away. Karaban does Karaban things (14/6/3 on efficient shooting) and Mullins provides a spark with two threes, but it's Reed who keeps the Huskies in this.

UConn's fatal flaw is what kills the upset. Solo Ball's three-point slump (1-for-5, extending a season-long regression from 41.4% to 29.6%) compresses UConn's spacing and lets Michigan's help defenders crowd the paint against Reed. Demary is outstanding defensively and fills the stat sheet (10/5/7 with 2 steals), but his three turnovers against Michigan's length — particularly Mara's ability to deflect passes in the lane — are costly. The Huskies go 13-for-20 from the stripe, and a team ranked 306th in free-throw rate can't afford to leave five points on the table in a five-point game.

🎯 UConn vs. Michigan player props

See our experts' UConn vs. Michigan player props, including our best bets to win MVP.

Tarris Reed Jr. Over 9.5 rebounds

Reed has averaged 13.0 rebounds per game in the NCAA Tournament on 58.9% of available minutes. His offensive rebounding rate (13.9%) and defensive rebounding rate (24.5%) are both elite, and Michigan allows a 27.7% offensive rebounding rate that ranks only 54th nationally — good but not suffocating. I'm projecting 12 rebounds, which should clear any line set at his 8.8 season average. Reed is going to live on the glass against his former team.

Aday Mara Over 12.5 points

Mara scored a career-high 26 points against Arizona when Lendeborg went down, and Lendeborg's reduced mobility means Mara will again shoulder more offensive responsibility. UConn allows a 40.5% free-throw rate — 307th in the country — and Mara's 7-3 frame should get to the line repeatedly against Reed and Reibe. His season average of 12.2 points per game doesn't reflect the version of Mara that Michigan will need tonight. I'm projecting 16 points, which should clear most lines.

🧪 AI prediction methodology

This projection was built by feeding Claude (Anthropic's AI model) the following data inputs before generating the simulation:

• Statistical profiles: Full KenPom scouting reports for both teams, including four factors (eFG%, turnover rate, offensive rebounding rate, free-throw rate) on both sides of the ball, adjusted efficiency ratings, tempo, strength of schedule, and lineup data. Michigan enters as KenPom's No. 1 overall team with a +39.72 net rating; UConn ranks 9th overall.
• Player-level data: Per-game stats and advanced metrics (PER, true shooting percentage, usage rate, box plus/minus, win shares) for every rotation player on both rosters, sourced from Sports Reference.
• Tournament performance: Game-by-game results and stat lines from both teams' NCAA Tournament runs, including Michigan's historic run of five consecutive 90-point, double-digit-win games and UConn's five consecutive wins holding opponents under 66 points.
• Injury data: Full reporting on Yaxel Lendeborg's sprained MCL and rolled ankle, including MRI results, minutes played against Arizona (14), his postgame comments, and Dusty May's assessment of his likely mobility level.
• Efficiency benchmarking: SBR's championship efficiency trend analysis comparing both teams' profiles against historical national champions since 2002.
• Odds and market data: Current spread (Michigan -6.5), total (145.5), and moneyline (Michigan -285 / UConn +230).
• Historical context: UConn's 6-0 record in national championship games, Michigan's 1-6 record in title games, Hurley's 11-0 streak in the Sweet 16 or later, and Tarris Reed Jr.'s connection to Michigan as a former Wolverine.

The simulation projected approximately 65 possessions per team based on both teams' adjusted tempo (Michigan 71.1, ranked 22nd; UConn 64.7, ranked 319th) and the likelihood that UConn controls pace. Individual stat lines were allocated using each player's usage rate, shot distribution, and efficiency metrics, then adjusted for the opposing defense's profile at each level (rim, mid-range, three-point). Lendeborg's minutes and production were reduced based on his injury limitations.

This is a projection, not a guarantee. No model — AI or otherwise — can account for the full randomness of a 40-minute basketball game. Use this as one data point among many.

📺 How to watch UConn vs. Michigan: 2026 national championship

Follow all of our college basketball analysis and expert predictions for March Madness.

• Date: Monday, April 6
• Tipoff: 8:50 p.m. ET
• Location: Lucas Oil Stadium (Indianapolis)
• TV: TBS/truTV
• Streaming: HBO Max