NCAA 2026 Office Pool Lab

Office pool bracket lab / 10000 simulations

Michigan cuts down the net.

Michigan over Florida, 81-79.

1 Michigan 81
58.3% title game win probability
1 Florida 79

Deterministic bracket

Path to the title

Monte Carlo board

Who survives each round

First-round pressure

Upset board

Portfolio case study / office pool brief

What went into it

Project arc

Built for an office March Madness pool, packaged like a product demo.

The goal was practical: make sharper bracket picks for a friendly office pool without relying on gut feel alone. The portfolio version shows the full path from messy public basketball data to a clean, explorable decision tool.

Use case Office pool edge

Pick a bracket that balances chalk, title upside, and selective upset risk.

Audience Non-technical coworkers

Readable outputs first, model details available for anyone who wants to inspect the logic.

Portfolio angle End-to-end data product

Research, scraping, cleaning, modeling, simulation, UI design, and storytelling in one piece.

01 Source capture

NCAA NET, AP, Sports Reference, Bracket Matrix, official bracket, schedules, and boxscore snapshots.

02 Data cleanup

Normalized team names, resolved source mismatches, merged field context, and tracked fallback quality.

03 Feature engineering

Recent-form windows, power ratings, seed priors, efficiency edges, pace, shooting, rebounding, and turnover matchups.

04 Historical training

Built tournament matchup rows, trained MLX logistic models, and compared a tree-based benchmark.

05 Production ensemble

Blended matchup model, tree benchmark, heuristic strength, and historical seed priors for calibrated picks.

06 Interactive demo

Packaged office-pool picks, round-reach odds, upset risk, data quality, and team search into a static UI.

Model room

How the pick was built

Ensemble weights

Validation

Data quality

Feature families

Field explorer

All 68 teams

Team Seed Conf Record NET SRS Final Four Champion