The Quant doesn't care about your alma mater. It cares about which outcome is most probable.
How It Works
Monte Carlo simulation: run the entire 63-game tournament 10,000 times, each time drawing outcomes from probability distributions. You can't enumerate every possible bracket (2^63 is about 9.2 quintillion). But you can simulate enough to see what keeps happening.
Each game's win probability comes from three inputs per team:
- Adjusted Offensive Efficiency -- points per 100 possessions, opponent-adjusted
- Adjusted Defensive Efficiency -- points allowed per 100 possessions, opponent-adjusted
- Adjusted Tempo -- possessions per 40 minutes
These come from KenPom data through conference tournaments. The efficiency gap between two teams gets converted to a win probability via a logistic function calibrated against 15 years of tournament results (2010-2025).
What "Most Probable" Means
After 10,000 runs, the model picks the team that won each individual game most often -- not the single most common complete bracket. The result is the highest-probability path through the tournament, game by game.
In practice, the champion typically appears in 20-35% of simulations. That's the reality of March Madness -- even the best team rarely has better than 1-in-3 odds.
Why This Beats Seeds
Seeds are a committee's subjective ranking. Efficiency ratings are performance data. An 8-seed can be statistically better than a 5-seed, and The Quant picks the upset accordingly. Over 15 years of backtesting, efficiency-based models outperform seed-based picking by roughly 8%.
Honest Limitations
No style-of-play mismatches (a slow team disrupting a fast team's rhythm). No injury data. No conference tournament momentum. Pure numbers, for better or worse.
Picks locked March 19. See how it holds up on the dashboard.