♟️ Engine Overview

This Python-based chess AI engine implements an advanced move evaluation system using a combination of traditional techniques and optimization heuristics such as Zobrist hashing, transposition tables, aspiration windows, Late Move Reductions, and Quiescence Search. Below is a breakdown of its main components:

🔢 Evaluation Components

• pieceScore: Defines base material values for each piece.

• Positional Score Tables (knightScores, bishopScores, etc.): Provide bonuses/penalties based on piece positions to guide piece activity and board control.

• piecePositionScores: Maps piece types to their respective position tables.

♻️ Zobrist Hashing & Transposition Table

• ZOBRIST_TABLE: Generates unique 64-bit hashes for board positions to enable fast position recognition.

• initializeZobrist(): Initializes the hash keys for all piece-square combinations and the current side to move.

• computeZobristHash(board, whiteToMove): Produces the Zobrist hash for a given board state.

• transpositionTable / TTEntry: Caches previously evaluated positions to avoid redundant computation and improve search speed.

🧠 Search Algorithms

• findBestMove(): Entry point for finding the optimal move using iterative deepening and aspiration windows to efficiently narrow the alpha-beta bounds.

• findMoveNegaScout(): Core NegaScout/PVS (Principal Variation Search) algorithm enhanced with:

  • Late Move Reduction (LMR)

  • Move ordering

  • History heuristic

  • Transposition table lookup and storage

• quiescenceSearch(): Prevents horizon effects by extending the search only through capture sequences.

• see() (Static Exchange Evaluation): Fast estimation of capture profitability without full search.

📋 Move Ordering

• orderMoves(): Scores moves based on captures, promotions, and historical success to guide search toward promising branches early.

• orderMovesWithBestFirst(): Keeps previously best move at the front for more stable iterative deepening.

📈 Evaluation Function

• evaluateBoard(): Combines material, positional, structural, and strategic factors:

  • Material balance

  • Positional tables

  • King safety

  • Pawn structure (isolated, doubled, passed pawns)

  • Promotion threats

  • Open files for rooks

  • Bishop pair bonus

• Supporting functions:

  • evaluateKingSafety(gs, color)

  • evaluatePawnStructure(gs, color)

  • evaluatePromotionThreats(gs)

📦 Utilities

• findRandomMoves(validMoves): Picks a random legal move (used as fallback or for testing).

• canAttackSquare(): Checks if a piece can attack a target square (partial legality check).

• isClearDiagonal() / isClearStraight(): Check if a piece’s path to a target square is clear.

🧪 Customization & Parameters

• DEPTH: Controls the depth of the main search (default: 5)

• CHECKMATE, STALEMATE: Constants used in scoring terminal game states.

• QUIESCENCE_DEPTH_LIMIT: Limit for depth extension during quiescence search.

🚀 Highlights

• Iterative Deepening for better time control and progressive deepening.

• Aspiration Windows for focused alpha-beta pruning.

• Zobrist Hashing enables fast state comparison.

• Static Exchange Evaluation (SEE) for smarter capture pruning.

• Quiescence Search for tactical stability near leaf nodes.

• Heuristics like history tables, position scoring, and pawn structure to emulate human evaluation principles.