Fantasy Football Draft Algorithm Comparison

A comprehensive comparison between traditional ADP-based (Greedy) and advanced Regret-based draft algorithms for fantasy football.

Project Overview

This project implements and compares two fantasy football draft algorithms:

1. Greedy Algorithm: Traditional approach using Average Draft Position (ADP) rankings
2. Regret Algorithm: Advanced approach considering positional scarcity, value dropoff, and pick timing

Quick Start

Prerequisites

pip install pandas numpy matplotlib seaborn openpyxl

Required Data Files

Ensure these files are in the data folder:

ESPN Draft Rankings:

• fantasy_rankings_2020.xlsx
• fantasy_rankings_2021.xlsx
• fantasy_rankings_2022.xlsx
• fantasy_rankings_2023.xlsx
• fantasy_rankings_2024.xlsx

Actual Season Performance:

• 2020.xlsx
• 2021.xlsx
• 2022.xlsx
• 2023.xlsx
• 2024.xlsx

Running the Analysis

Run complete analysis

python main.py

Run tests

python test_suite.py

Test specific algorithm

python -c "from main import *; # your custom test code"

File Structure

├── main.py # Main execution file ├── models.py # Data models (Player, Team) ├── algorithms.py # Draft algorithms implementation ├── data_loader.py # Data loading utilities ├── draft_simulator.py # Draft simulation engine ├── performance_analyzer.py # Performance benchmarking ├── visualizer.py # Chart generation ├── test_suite.py # Automated tests └── README.md # This file

Algorithm Details

Greedy Algorithm

• Time Complexity: O(n log n)
• Space Complexity: O(n)
• Strategy: Pick best available player by ADP for needed positions

Regret Algorithm

• Time Complexity: O(n²) per pick
• Space Complexity: O(n)
• Strategy: Calculate regret scores based on:
  • Player value (normalized auction value)
  • Positional scarcity (quality players remaining)
  • Value dropoff (gap to next best player)
  • Pick urgency (time until next turn)

Regret Score Formula:

regret = value_score + (scarcity × dropoff × 0.4) + (urgency × 0.3) - bye_penalty

Draft Configuration

• Teams: 10 (snake draft)
• Draft Position: 5th
• Roster: 1 QB, 2 RB, 2 WR, 1 TE, 1 FLEX, 1 K
• Total Picks: 8 per team

Contributing

To extend the algorithm:

1. Add new factors to regret score calculation
2. Implement alternative draft strategies
3. Add new performance metrics
4. Expand test coverage