Logo Lanfrica

drwjkirkpatrick-web/giuthi-ai-explorer

Domaine:

education

Type de record:

software
Créateur:
drw
Hôte:
Interactive AI visualization tool for Giuthi, a traditional Kikuyu mancala game from Kenya — teaches heuristic evaluation, alpha-beta pruning, and relay sowing to ages 10-17 # Giuthi — Interactive AI Explorer 🇰🇪 An educational interactive tool for understanding how game AI works, built around **Giuthi**, a traditional Kikuyu mancala board game from central Kenya. ## What Is Giuthi? Giuthi (pronounced *gee-OO-thee*) is a mancala game played by the Kikuyu people of Kenya. The name means "to move" or "to transfer" in Kikuyu. It belongs to the ancient mancala family of games, played across Africa for over 1,000 years. ### How to Play - **Board**: Two rows of pits (default 6 per row), with seeds in each pit - **Goal**: Capture more seeds than your opponent - **Sowing**: Pick up all seeds from one of your pits and drop them one-by-one into subsequent pits — clockwise or counterclockwise (your choice) - **Relay sowing**: If your last seed lands in a non-empty pit, pick up all seeds there and continue in the **opposite direction**. This can chain multiple times! - **Capture**: If your last seed lands in an **empty pit on your side** and you sowed into enemy territory, capture the seeds in the opposite pit - **Chain captures**: After capturing, check the next pit — if it's also empty on your side with seeds opposite, capture those too - **Starting rule**: You can only play from a pit with 2+ seeds (singletons can't be played) - **Game end**: When neither player can move. Most seeds wins! ## What Makes This Different from Shisima? This is the second game in a progression teaching game AI concepts: | | Shisima | Giuthi | |---|---|---| | **Game type** | Alignment (3-in-a-row) | Mancala (sowing + capture) | | **State size** | ~3,360 positions | 10¹⁵+ positions | | **AI search** | Full game tree | Partial search + heuristic | | **Key lesson** | Minimax basics | Heuristic evaluation + pruning visualization | | **Repo** | shisima-game-lesson | This repo | ## What Students Learn - **Heuristic evaluation functions** — how AI judges positions it can't fully analyze - **Alpha-beta pruning** — visualized in real-time, showing *why* pruning matters …