My planned phases for turning the documented rules into a working digital game with AI opponents.
Organize rules, glossary, setup, pieces, FAQ decisions, and known ambiguities. This HTML archive is the starting point.
Create Python data structures for board coordinates, players, pieces, zones, Gyullas, compass states, graveyard, keep, and turn state.
Implement Gulled movement, Navia movement, Maseitai compass movement, captures, summoning, line-over, Dratp/Invoke timing, and check detection.
Build scenario-based unit tests for every rule. No UI work should bypass the engine.
Start with CLI or simple Pygame board view. The UI should call the same engine used by tests and AI.
Begin with random legal move AI, then heuristic evaluation, then minimax/alpha-beta, and eventually a learning or Monte Carlo approach if useful.
Add piece art, gallery references, tutorial mode, save/load, move history, replay export, and configurable armies.
