A high-performance chess engine built in Rust, featuring a Negamax search with Alpha-Beta pruning and a Tapered Evaluation system. I have structured this project to utilize the library/binary split effectively, as well as manual trait implementations, external data processing, and command-line integration.
- Search Algorithm: Implements Negamax with Alpha-Beta Pruning to navigate the search tree efficiently.
- Quiescence Search: Extends the search for captures and promotions until a stable board state is reached.
- Tapered Evaluation: Linearly interpolates between Midgame and Endgame scoring based on the remaining material on the board.
- Positional Awareness: Uses PeSTO Piece-Square Tables (deserialized via
serdefrom JSON) to evaluate piece placement. - User-Friendly CLI: Powered by
clap, allowing for easy FEN input and depth control.
The line coverage as per llvm-cov is listed here.
Provide a FEN string and an optional depth (default is 5).
The sample provided here is an 1901 rated Lichess puzzle, in which at depth 6 the engine provides the optimal move and verifies that we have entered a mating sequence (which involves a queen sacrifice for line completion).
cargo run -- --depth 6 "1r5r/1ppk3q/p7/N1nP4/P4p2/1R4N1/1P2QPP1/R5K1 b - - 0 31"