Scrabble

Quackle_Self_Play is a data analysis project that leverages the Scrabble AI Quackle to simulate 250,000 self-play games in Speedy Player mode. It records every word played during these matches, including hooks and repeats, using the OWL3 dictionary. The project provides insights into AI-driven Scrabble strategy by analyzing word frequency and tile usage patterns.

Key Features

• Self-Play Simulation — Quackle AI plays against itself 250,000 times to generate comprehensive game data.
• Word Tracking — Records all words played, including hooks (e.g., extending ULLAGES to SULLAGES).
• Power Tile Analysis — Examines usage patterns of high-value tiles (J, Q, X, Z).
• Word Length Statistics — Identifies most frequently used 2, 3, 4, and 5-letter words.
• Raw Data Output — Provides unprocessed text files of all recorded words for further analysis.
• Dictionary Compatibility — Uses OWL3 with plans to update to OWL4 upon release.

Philosophy

The project emphasizes data-driven exploration of Scrabble AI behavior, offering transparent, unmodified datasets to support strategy analysis and linguistic research.