Chess Game Accuracy Analysis Script

This script processes chess PGN (Portable Game Notation) data to compute the average centipawn loss and a mean of volatility-weighted mean and harmonic mean (Accuracy) for both players.

Requirements

• Python 3.6 or higher
• chess Python library
• Stockfish chess engine (compatible version for your system)

Installation

1. Install Python 3.6+ from python.org.
2. Install the chess library using pip:

   pip install python-chess

3. Download the Stockfish engine from the official Stockfish website and place it in a known directory.

Usage

Basic usage requires specifying the analysis depth, number of threads, and path to the Stockfish engine. You can input PGN data either from a file or directly as a string:

From a File

python chess_accuracy.py 16 2 /path/to/stockfish -file=/path/to/game.pgn

From a PGN String

python chess_accuracy.py 16 2 /path/to/stockfish -pgn="1.e4 e5 2.Nf3 Nc6 3.Bb5 a6"

Using Standard Input

You can also pipe PGN data into the script:

cat /path/to/game.pgn | python chess_accuracy.py 18 2 /path/to/stockfish

Verbose Output

For detailed output, including move-by-move analysis, add the -verbose flag at the end:

python chess_accuracy.py 16 2 /path/to/stockfish -file=/path/to/game.pgn -verbose

Output

The script outputs the average centipawn loss and harmonic mean of accuracy for both players, formatted as follows:

Average centipawn loss (White), Accuracy (White), Average centipawn loss (Black), Accuracy (Black):
44, 84, 15, 95

For verbose mode, additional details for each move are printed to the console.

License

MIT