This project will be deleted soon!

Due to GitHub's unfortunate new practice of using repositories to train Microsoft LLMs, I have made the difficult decision to migrate all of my projects to Codeberg.

Because of this project's low visibility, I have decided that it is not important enough to keep on GitHub. Therefore, I have decided to delete this project at some point during the month of September. This project's new home can be found here.

Build

The CMake build scripts are only tested on Linux using the GNU C++ compiler (for native compilation) and the MinGW C++ compiler (for cross-compilation to Windows). MacOS (with Apple Silicon) seems to work in my limited testing, since g++ on MacOS is just a link to the LLVM C++ compiler.

To be honest, I have never used CMake on Windows, so I don't even know how to start with that.

# Clone repo and submodules
git clone https://github.com/computablee/100-Elo-Chess-Engine.git
cd 100-Elo-Chess-Engine
git submodule init
git submodule update

# Build
mkdir build
cd build
cmake .. # -DWINDOWS=true to cross-compile to Windows
make     # -jN for multi-core compilation

Current Elo Estimates

Update: The 100 Elo Chess Engine has been released to CCRL for evaluation. This README will be updated with a rating as soon as I'm aware of it.

Currently, the 100 Elo Chess Engine is using the Stash engine and cutechess to get a rough idea of Elo. We have compared to:

• v9 (1287 Elo)
• v11 (1698 Elo)
• v12 (1891 Elo)
• v13 (1977 Elo)
• v14 (2068 Elo)
• v15 (estimated 2150 Elo)

Latest run was done against v15 with a time control of 120+1 (in seconds) with 1000 total runs.

Results may be outdated. I run these occasionally; the engine may be stronger than the results shown.

Score of 100 Elo Chess Engine vs Stash v15.0: 331 - 275 - 394  [0.528] 1000
...      100 Elo Chess Engine playing White: 181 - 114 - 205  [0.567] 500
...      100 Elo Chess Engine playing Black: 150 - 161 - 189  [0.489] 500
...      White vs Black: 342 - 264 - 394  [0.539] 1000
Elo difference: 19.5 +/- 16.8, LOS: 98.9 %, DrawRatio: 39.4 %

This results in an estimated Elo of: 2170.

Current Features

The 100 Elo Chess Engine uses the following architecture:

• Basic/buggy UCI support
• PeSTO with incremental evaluation
• Iterative deepening
• Transposition table
• Null move pruning
• Move ordering (best move -> MVV-LVA -> killer moves -> history heuristic)
• Quiescent search
• Principal variation search
• Late move reduction
• Check extensions
• Internal iterative reductions
• Reverse futility pruning