ML Trading System

An automated Python-based machine learning trading system for futures contracts.

Overview

This project implements a sophisticated, modular trading system that operates locally while ensuring robust risk management. The system collects minute-level market data, researches and backtests both traditional and machine learning-driven strategies, and automatically executes trades when predefined risk and performance criteria are met.

Getting Started

Prerequisites

• Python 3.8+
• TradingView API access
• Tradeovate account (for live trading)

Installation & Usage

For detailed instructions on how to install, configure, and run the system, please refer to the Run Instructions document.

Features

Data Collection

• Minute-level market data from TradingView API
• Robust data cleaning and normalization
• Efficient local storage with S3 bucket support

Trading Strategies

• Traditional Strategies:

  • Moving Average Crossover
  • Bollinger Bands
  • VWAP (Volume-Weighted Average Price)
  • MACD

• Machine Learning Strategies:

  • Random Forest
  • XGBoost
  • Neural Networks

• Advanced Technical Indicators:

  • VWAP (Volume-Weighted Average Price)
  • Pivot Points
  • Trendline Breakouts
  • Bull and Bear Flag Patterns

Backtesting

• Event-driven backtesting engine
• Dynamic slippage modeling
• Comprehensive performance metrics

Risk Management

• Win rate threshold (minimum 51%)
• Risk-to-reward ratio requirements (1:2 or 1:3)
• Maximum daily loss limits ($2,500)
• Configurable position sizing

Project Structure

• src/ - Main source code
  • data/ - Data collection and processing
  • strategies/ - Trading strategies
  • backtesting/ - Backtesting engine
  • execution/ - Trade execution
  • logs/ - Logging and monitoring
• config/ - Configuration files
• tests/ - Unit and integration tests

Risk Management

The system enforces strict risk management rules:

• Only executes trades for strategies with at least 51% win rate
• Requires a minimum risk-to-reward ratio of 1:2 (preferably 1:3)
• Enforces a maximum daily loss limit of $2,500

Features

• Data collection from TradingView API
• Strategy research and backtesting
• ML-based strategy discovery
• Automated trade execution
• Risk management with configurable parameters
• Comprehensive logging and monitoring

Requirements

• Python 3.11.4 or higher
• Virtual environment for dependency management

Setup

1. Clone this repository
2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Configure your API keys in config/config.yaml
5. Run the application:

   python main.py
   

Project Structure

• src/ - Main source code
  • data/ - Data collection and processing
  • strategies/ - Trading strategies
  • backtesting/ - Backtesting engine
  • execution/ - Trade execution
  • logs/ - Logging and monitoring
• config/ - Configuration files
• tests/ - Unit and integration tests

Risk Management

The system enforces strict risk management rules:

• Minimum win rate: 51%
• Minimum risk-to-reward ratio: 1:2 (preferably 1:3)
• Maximum daily loss: $2,500

Future Enhancements

• Migration to AWS with S3 bucket storage
• Web-based dashboard for monitoring
• Advanced performance analytics