march-madness - SKILL.md Agent Skill

name: march-madness description: NCAA college basketball analytics and March Madness prediction tools. Use when analyzing college basketball games, finding betting edges, building bracket predictions, or accessing NCAAB data. Includes tools for CBBpy (ESPN data), Barttorvik T-Rank, DRatings predictions, NCAA NET rankings, Massey composite ratings, Warren Nolan metrics, Kaggle tournament data, Haslametrics, StatMuse (natural language queries), Sports-Reference.com, SportsData.io API, FantasyPros projections, and multi-source odds comparison. All data sources are FREE.

March Madness NCAA Basketball Analytics

A comprehensive skill for NCAA college basketball data analysis and March Madness prediction modeling.

Features

🏀 Complete Data Coverage: 25+ free data sources for NCAA basketball analytics 📊 Multiple Prediction Models: Access to T-Rank, BPI, NET, KenPom-style metrics, and computer ratings 💰 Edge Finding: Compare model predictions vs betting lines for profitable opportunities 📈 Historical Data: Multi-season data for backtesting and model training 🎯 Tournament Focus: Specialized tools for March Madness bracket prediction

Available Tools

Core Data Tools

cbbpy_tools.py: ESPN data via CBBpy (schedules, boxscores, play-by-play, player stats)
espn_scraper.py: Direct ESPN API access (BPI ratings, scores, standings)
ncaa_stats.py: Official NCAA statistics and NET rankings
kaggle_data.py: March Madness historical datasets

Analytics Sites

barttorvik.py: T-Rank ratings and tempo-free analytics
warren_nolan.py: NET rankings, RPI, strength of schedule
massey_ratings.py: Composite computer ratings from multiple systems
haslametrics.py: Computer predictions and unique analytics

New Sources (Derek's picks)

statmuse.py: Natural language basketball queries (ask anything in plain English)
sportsreference.py: Sports-Reference.com team/player stats, rankings, schedules
sportsdata_io.py: SportsData.io REST API (scores, odds, projections, stats)
fantasypros.py: FantasyPros projections, expert picks, NCAAB news

Betting & Predictions

dratings.py: Computer predictions with odds comparison
odds_comparison.py: Multi-source odds and trends
edge_finder.py: Identify profitable betting opportunities

Quick Start

# Get today's games with predictions
python3 scripts/cbbpy_tools.py games-today --with-odds

# Find betting edges for today's games
python3 scripts/edge_finder.py today --min-edge 3.0

# Get T-Rank ratings for top 25 teams
python3 scripts/barttorvik.py rankings --top 25

# Compare model predictions vs betting lines
python3 scripts/dratings.py edge-finder --date today

Installation

cd skills/march-madness
pip3 install -r requirements.txt

Data Sources

All data sources are FREE and documented in references/data_sources.md. Key sources include:

CBBpy: ESPN play-by-play and boxscore data
Barttorvik: T-Rank tempo-free analytics
NCAA: Official NET rankings and statistics
DRatings: Computer predictions and betting analysis
Kaggle: Historical March Madness tournament data
Warren Nolan: Comprehensive NCAA metrics tracking

Usage Examples

Daily Game Analysis

# Get today's schedule with predictions
python3 scripts/cbbpy_tools.py games-today

# Get current T-Rank top 25
python3 scripts/barttorvik.py rankings --top 25

# Find today's best betting edges
python3 scripts/edge_finder.py scan --date today

Team Research

# Get Duke's full schedule and stats
python3 scripts/cbbpy_tools.py schedule --team "Duke" --season 2026

# Get Duke's T-Rank metrics and predictions
python3 scripts/barttorvik.py team-stats --team "Duke"

# Get Duke's NET ranking and quad records
python3 scripts/warren_nolan.py net --team "Duke"

Bracket Building

# Get current bracket projections
python3 scripts/warren_nolan.py bracket-projection

# Get computer rating consensus
python3 scripts/massey_ratings.py composite --top 68

# Backtest model performance
python3 scripts/edge_finder.py backtest --start-date 2025-03-01

Output Formats

All tools support:

JSON: Default machine-readable format
Pretty: Human-readable format with --pretty flag
CSV: Structured data export with --csv flag

References

references/data_sources.md: Complete documentation of all 25 data sources
references/metrics_guide.md: Guide to understanding advanced basketball metrics
references/march_madness_calendar.md: Important dates and tournament structure

Contributing

This skill is open source. Contributions welcome for additional data sources, prediction models, or analytics tools.

License

MIT License - see LICENSE file for details.