Project Patti: Why can You Solve Diabolical Puzzles on one Sudoku Website but not Easy Puzzles on another Sudoku Website?
2025-07-30
Summary
The paper investigates the varying difficulty ratings of Sudoku puzzles across different websites by proposing two new metrics: Clause Length Distribution derived from SAT (Satisfiability) encodings, and Nishio Human Cycles, which simulate human Sudoku-solving strategies. The study analyzes over a thousand puzzles from five websites and finds that while four sites show strong correlations between the proposed metrics and their labeled difficulty levels, one site is inconsistent. A universal difficulty classification system is developed to categorize puzzles into Universal Easy, Medium, and Hard levels.
Why This Matters
Understanding what makes Sudoku puzzles difficult can help standardize difficulty ratings across platforms, improving user experience and providing more consistent challenges to players. The development of a universal difficulty classification allows for easier comparison of puzzles from different sources, potentially aiding educators or developers who rely on consistent difficulty levels for instructional or entertainment purposes.
How You Can Use This Info
Professionals involved in game design or education can use these metrics to more accurately assess and standardize puzzle difficulties across platforms, ensuring a consistent user experience. Additionally, the insights from this study could be applied to other problem-solving contexts where difficulty assessment is crucial. For those involved in training AI models, the methods explored could inform approaches to simulate human problem-solving behavior.