SoK Introductory Page

Season of KDE 2025 – Bringing Pallanguli to Mankala Engine

Hello everyone! I’m excited to be a part of Season of KDE 2025, continuing my contributions to the Mankala Engine by integrating the Pallanguli variant of Mancala.

Pallanguli is a fascinating two-player game with deep strategic elements and a rich cultural history in South India. Unlike many Mancala variants, it features long, cascading moves, intricate capture mechanics, and a dynamic board state that makes for highly engaging gameplay. Implementing Pallanguli within the Mankala Engine is both a technical and theoretical challenge—one that I’m eager to take on! Enhancing AI & Game Strategy

While integrating the game’s mechanics, I’m also focusing on improving AI strength and decision-making. The current engine uses Minimax with Alpha-Beta Pruning, which is effective but may not fully capture the complexity of Pallanguli’s branching game tree. That’s why I’m exploring Monte Carlo Tree Search (MCTS) and other heuristic-driven approaches to see if they offer better performance.

Additionally, I plan to build an opening book for Pallanguli, similar to those used in chess engines. By precomputing strong opening sequences, the AI can make optimal early-game decisions, reducing the need for brute-force searching while maintaining strong play. This will not only make the AI more efficient but also provide a structured way for players to learn and understand strong opening moves. The Road Ahead

There’s a lot to do, from refining the move generation and evaluation functions to ensuring smooth integration with the existing engine. I’m also excited to collaborate with the KDE community, get feedback, and improve the engine’s usability for both casual players and AI researchers interested in strategic games.

Looking forward to an exciting Season of KDE, learning more, and making Mankala Engine a better platform for exploring Mancala variants! 🚀