▄▄    ▄▄         ▄▄                                                        
                     ██  ▀███       ▀███                                   ██                   
                           ██         ██                                   ██                   
 ▄█▀████████  ▀███ ▀███    ██    ▄█▀▀███ ▀████████▄█████▄  ▄█▀██▄  ▄██▀████████  ▄▄█▀██▀███▄███ 
▄██  ██   ██    ██   ██    ██  ▄██    ██   ██    ██    ██ ██   ██  ██   ▀▀ ██   ▄█▀   ██ ██▀ ▀▀ 
▀█████▀   ██    ██   ██    ██  ███    ██   ██    ██    ██  ▄█████  ▀█████▄ ██   ██▀▀▀▀▀▀ ██     
██        ██    ██   ██    ██  ▀██    ██   ██    ██    ██ ██   ██  █▄   ██ ██   ██▄    ▄ ██     
 ███████  ▀████▀███▄████▄▄████▄ ▀████▀███▄████  ████  ████▄████▀██▄██████▀ ▀████ ▀█████▀████▄   
█▀     ██                                                                                       
██████▀                                                                                         

Guildmaster is a lightweight Go application that calculates and tracks player rankings using the Elo rating system. The tool reads game results (particularly Magic: The Gathering games) from CSV data and computes updated ratings using a pairwise Elo scoring method suitable for multiplayer ranked games.

• Elo Rating Calculation: Implements the Elo rating system to provide an objective measure of player skill
• CSV Data Processing: Reads game records from a CSV file to calculate ratings
• Player Performance Tracking: Maintains and updates player ratings based on game outcomes
• Flexible Input: Supports a variety of game formats where players are ranked from winner to losers
• Terminal User Interface: View rankings in an interactive TUI with navigation controls

Guildmaster scores multiplayer games by treating each finished game as a set of pairwise matches derived from the final ranking. For a game with N players, every higher-placed player is considered to have "won" against every lower-placed player in that same game. Examples:

• 1st place beats 2nd, 3rd, 4th, ...
• 2nd place loses to 1st but beats 3rd, 4th, ...
• 3rd place loses to 1st and 2nd but beats 4th, ...

All pairwise expected scores are computed using Elo with the configured D constant and K factor. By default Guildmaster uses:

• K = 40
• D = 800

Updates are calculated from a snapshot of player ratings before the game, accumulated as deltas for each player (summing every pairwise result), and applied once per player. This prevents order-dependent updates and ensures deterministic results when combined with the stable sorting tie-breaker (player name) used when presenting rankings.

This approach gives more consistent and interpretable rating changes for multiplayer events compared to only scoring adjacent placements.

• Go 1.25
• Dependencies:
  • github.com/kortemy/elo-go
  • Make

# Clone the repository and navigate to that directory
git clone https://github.com/dylanlott/guildmaster.git
cd guildmaster

# Build the application locally 
make build

# Run the scoring algorith
make run

# Run the terminal application 
make tui

1. Prepare your game data in a CSV file with the following format:

   • First column: Empty (or can contain identifier)
   • Second column: Date of the game
   • Subsequent columns: Players in order of finish (winner first)

2. Run the application:

# Using the default CSV file (mtgscores.csv)
./guildmaster

# Using a custom CSV file
./guildmaster -path=path/to/your/data.csv

# Display rankings with Terminal User Interface
./guildmaster -tui

The -tui flag enables an interactive Terminal User Interface for viewing player rankings:

• Navigate through the rankings using the up and down arrow keys
• Exit the TUI by pressing q or Ctrl+C

This provides a more interactive way to browse player rankings, especially when dealing with large player pools.

Run the app in TUI mode:

go run . -tui

The expected CSV format looks like:

,2022-01-01,PlayerA,PlayerB,PlayerC
,2022-01-08,PlayerB,PlayerA,PlayerD

Where:

• The first column is empty
• The second column contains the date
• The remaining columns list players in finishing order (winner to losers)

A small HTTP server is provided under cmd/server. It exposes a tiny REST API and serves static files from the assets/ directory.

Endpoints:

• GET /api/scores -> returns current scores as JSON

• POST /api/game -> accepts {"players": ["A","B",...]}, computes Elo deltas and persists them in-memory

Run the server locally:

# from the project root
make server
# or
go run ./cmd/server

Open http://localhost:8080 to view the minimal web UI (assets/index.html).

The application:

1. Reads the game data from the specified CSV file
2. Initializes each player with a base Elo rating
3. For each game, updates player ratings based on their performance using pairwise outcomes
4. Calculates and displays the final Elo ratings for all players

Contributions are welcome! Please feel free to submit a Pull Request.

This project is licensed under the MIT License - see the LICENSE file for details.