Better than Bronze, Worse than Gold. A very humble attempt at a chess engine written in pure C.
So far, the engine is capable of generating all legal moves for a certain player, given a board position. This means no automation, no AI, no robust decision making. Yet. Don't give up on me, pls.
The implementation, however, is modular enough to easily allow for checkmate/stalemate/draw detection in any end-game scenario, with little more code (it is in fact my next goal). Further on I plan on implementing a search mechanism based on alpha-beta search and iterative deepening, and a static evaluation mechanism with hopefully-not-so-obvious heuristics, to improve move scoring and the still-non-existing decision making ability.
Right now it's closer to a cheese engine, but I plan to make it fully fledged. I intend to make it UCI compliant to be able to pit it against other engines in automated contests, when it's minimally ready. I'll also be collecting Perft scores for my move generation mechanism, to understand better what performance means in this context.
Provide the program with a FEN-compliant string and it will list all the possible moves for the current player in algebraic notation, ordered by score (best to worst), along with a drawn board and the Zobrist hash code for the position. The scoring mechanism is, so far, simple: It takes into account the moving piece with big bonuses for captures and promotions, varying with piece captured or promoted to. As I said, there's no evaluation mechanism yet to provide an acceptably precise score for each move.
Here's an example input...
rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1
...And it's following output
A B C D E F G H
*-----------------*
8 | r n b q k b n r | 8
7 | p p p p p p p p | 7
6 | . . . . . . . . | 6
5 | . . . . . . . . | 5
4 | . . . . . . . . | 4
3 | . . . . . . . . | 3
2 | P P P P P P P P | 2
1 | R N B Q K B N R | 1
*-----------------*
A B C D E F G H
Zobrist hash key for position: 8748045942876076066
h3
h4
g3
g4
f3
f4
e3
e4
d3
d4
c3
c4
b3
b4
a3
a4
Nh3
Nf3
Nc3
Na3
You'll need CMake and there are pre-compiled binaries available for whatever platform you're on. If you're on Unix, you can use your package manager to install it (that includes brew, on Mac). If you're on Windows, I suggest setting up an environment with CygWin and here's a good reason why.
Standard CMake build: do everything in a separate build dir or similar, to
keep things clean. From the project's root dir, you can go something like this:
mkdir build && cd build
cmake -DCMAKE_C_COMPILER=gcc -DCMAKE_EXPORT_COMPILE_COMMANDS=on -DCMAKE_BUILD_TYPE=Release ..
This is what a typical cmake command I issue looks like, mostly all preference.
I don't have a Release target yet and Debug is not only useful but still
very necessary. Then you can just make and run the binary:
make
./ChessEngine
Warnings will follow. This is still very barebone-ish, but I'm gettin' there.
One primarily does a thing like this to learn it, but there is no greater satisfaction than inspiring other people. If that's the case, I can help...
- First, I suggest you read pretty much all you can from Bruce Moreland's notes on Chess Programming and the Chess Programming Wiki, these were my main go-to resources for acquiring knowledge into chess programming. For theory and application as well.
- Then, the CPW Engine is a good read to leverage the code-writing process, if you have an idea of what you're doing/want to do (the code is pretty not-so-pretty and poorly commented, I'm afraid, since play power and low resource usage seem to be the main goals).
- Tom Kerrigan's TSCP is, on the other hand, very well commented. I've acquired some pretty good insights reading it's code. For such a simple code (the focus is readability), the play power is very respectable.
If you think this project to have any potential of becoming something worth a glance then you're more than welcome to contribute to it.