Chess programming

I've been fascinated by computer chess since I was a child.

One Christmas I was given a Chess Challenger, and it took me ages before I could beat it. As soon as I started programming I was trying to write chess programs of my own, and I recognised early on that searching the tree of possible moves needed a recursive approach. There were very few books or magazines about chess programming then, and no internet, so I worked it out slowly for myself, rediscovering minimax, negamax and alpha-beta along the way.

My first real chess program, written when I was fifteen or sixteen, was in Z80 assembly language on a 16K ZX Spectrum. It played terribly, but it played legal chess, promotions, castling and en passant included. I kept improving it, but soon realised that assembly language was too hard going, so my later engines have been written in C and then C++. More recently I've enjoyed porting them to Rust and Go, two languages I really like.

Computer chess has changed enormously over the years, and I've learnt along with it: the bitboard board representation, the many ways to prune and speed up the search, evaluation functions tuned by heuristics, by hand and through games, and most recently how to build and train efficiently updatable neural networks (NNUE) to evaluate positions.

Everything here is open source under the GNU General Public License. I hope you find the code easy to read and understand, well documented, and strong over the board.