Loading...
Champion--
Avg Conf--
Upsets--
Leverage--
Updated--
ROUND 1Mar 19-20
ROUND 2Mar 21-22
SWEET 16Mar 26-27
ELITE 8Mar 28-29
FINAL 4Apr 4-6
ELITE 8Mar 28-29
SWEET 16Mar 26-27
ROUND 2Mar 21-22
ROUND 1Mar 19-20
High confidence (70%+) Medium (50-70%) Low (<50%) Upset pick Leverage pick Name = Predicted winner
Loading analytics...
Loading backtest data...
ABOUT THIS PROJECT
The story behind the AI bracket
ABOUT THIS PROJECT
AI-powered March Madness bracket predictor built on March 18, 2026. Uses Claude (Anthropic), GPT-4o (OpenAI), and Gemini (Google) in a multi-model ensemble to predict every game from the Round of 64 through the National Championship.
HOW IT WORKS
1. DATA COLLECTION
KenPom ratings, BartTorvik T-Rank, NCAA NET rankings, injury reports, beat writer intel
2. MULTI-MODEL ENSEMBLE
Each matchup is analyzed by Claude, GPT-4o, and Gemini independently. Their predictions are compared using ensemble disagreement rules.
3. CONFIDENCE FORMULA
Weighted formula: 55% model strength, 20% source agreement, 15% lineup certainty, 10% data freshness
4. CROSS-AGENT DEBATE
The 10 most uncertain picks undergo a structured debate where the AI plays devil's advocate against its own predictions. 2 picks were flipped and average confidence dropped 8.4 points, reducing overconfidence.
5. MONTE CARLO SENSITIVITY
250 simulations testing different weight combinations against 2022-2025 historical results to find optimal calibration.
6. HISTORICAL BACKTESTING
Model accuracy scored against actual tournament results from 2022-2025, spanning ultra-chalk (2025) to extreme-chaos (2023) tournaments.
PREDICTION EVOLUTION
How the champion pick evolved through each stage of the pipeline
Initial run (Claude only)
UConn
After ensemble (Claude + GPT-4o)
Duke
After cross-agent debate
Duke (2 R64 picks flipped)
Final prediction
Duke over Houston predicted total score 152
TECH STACK
Python backend with Anthropic (Claude), OpenAI (GPT-4o), and Google (Gemini 2.5 Flash) SDKs. KenPom and BartTorvik for statistical data. Static HTML dashboard deployed on Vercel. Source code at github.com/elstonj/march-madness-2026.