♟️ Personal Chess Bot — Trained on My Own Games I was inspired by the idea of chess bots that mimic the styles of famous players on Chess.com. After seeing how unique and personalized those bots are, I decided to create my own chess bot that plays like me.

🧠 Inspiration & Idea While exploring Chess.com bots based on grandmasters and streamers, I wondered:

"What if I could train a bot to play like myself?"

So I exported my own games from Chess.com using Linktree, and began working on a bot that learns my move patterns and decision-making.

🔨 What I Did

✅ Data Collection: Downloaded my personal gameplay data from Chess.com using linkedtree

✅ Data Processing: Converted each game into FEN positions (board state) and corresponding best move responses

✅ Model Training: Used Random Forest Regression to predict my moves based on board positions

✅ Enhancement: Integrated Stockfish as a helper to improve move quality where predictions are uncertain

✅ Interface: Built a playable chess game GUI using Qt, where you can play against the trained bot

🚀 Features

Mimics your personal playstyle

Learns from real game data

Hybrid decision-making: combines ML predictions with Stockfish

Graphical interface for playing directly against your own AI

🧰 Tech Stack

Language: Python

ML Model: Random Forest Regression (via Scikit-learn)

Chess Engine: Stockfish

Data Format: PGN → FEN

GUI: PyQt5 / Qt for Python

♟️ Stockfish Setup

This project uses Stockfish to assist with move evaluation.

🔧 How to set it up:

1. Download the Stockfish binary for your OS from the official site.
2. Place the stockfish executable inside the project folder.
3. Make sure your Python script points to the correct file path:

   engine = chess.engine.SimpleEngine.popen_uci("stockfish")

📁 Coming Soon

Adjustable difficulty using model confidence

Side-by-side comparison with Stockfish-only bot