Chess

The engine started as a desktop game I wrote in C++ and was rewritten as a library that does one thing: it takes a position and answers with a move. The board is held as 64-bit bitboards, one bit per square for each kind of piece, so the legal moves of a side come out of a few shifts and table lookups. Sliding pieces use magic bitboards, which turn the question of which squares a rook can reach past the pieces in its way into one multiplication and one array read.¹

To choose a move it searches the tree of continuations with alpha-beta pruning,² which drops any line a player would never allow. Moves are tried most promising first: the hash move, then captures by the value of the victim, then the quiet moves that caused cutoffs before. Positions already seen are recognised by a Zobrist hash.³ Captures are played out at the horizon, so the search does not stop in the middle of an exchange, and it deepens one ply at a time, so a move is always ready when the clock runs out. Lines that are already good enough are cut short with a null move test, and late quiet moves are searched less deeply.¹ A position is scored by material and by where each piece stands, using piece-square tables.

The settings panel controls the search directly. Search depth is the most plies it looks ahead, time limit is the most it may think about one move, move hints marks the squares the piece you picked up can reach, and AI hints draws an arrow for the move the engine would play in your place. The charts under the board show how long each search took and how the engine rated the position after each of its moves.

Against the plain fixed-depth search it grew out of, the finished engine scored 96.3% over 200 games at 50 milliseconds per move. Against Stockfish 19 held to a chosen playing strength, at one second per move, it plays at about 2300 on Stockfish's own scale over 600 games.⁴ The source, the tests and every recorded benchmark step are in the repository.⁵

Footnotes

1. Chess Programming Wiki. Bitboards, magic bitboards, null move pruning and late move reductions. https://www.chessprogramming.org ²

2. Knuth, D. E., & Moore, R. W. (1975). An analysis of alpha-beta pruning. Artificial Intelligence, 6(4), 293-326. https://doi.org/10.1016/0004-3702(75)90019-3

3. Zobrist, A. L. (1970). A new hashing method with application for game playing (Technical Report 88). Computer Sciences Department, University of Wisconsin. http://digital.library.wisc.edu/1793/57624

4. Stockfish developers. Stockfish [Source code]. The strength is set with the UCI options UCI_LimitStrength and UCI_Elo. https://stockfishchess.org

5. Taha, S. chess [Source code]. GitHub. https://github.com/syedtaha22/chess