Quantitative Trading Portfolio SDK

Overview

This project provides a sophisticated SDK for evaluating and managing multiple trading strategies (robots) as a cohesive portfolio for the Brazilian stock market (B3). Unlike traditional trading platforms, our approach leverages high-resolution tick data and real-time portfolio analytics to optimize trade execution.

Purpose

Traditional backtesting platforms (like Strategy Quant) rely on simplified market approximations and tend to evaluate strategies in isolation. Our SDK addresses these limitations by:

1. Using accurate tick data instead of unreliable OHLC data from MT5
2. Evaluating multiple strategies together to understand portfolio-level dynamics
3. Optimizing trade exits based on collective performance metrics like Maximum Favorable Excursion (MFE)
4. Providing real-time portfolio evaluation for dynamic decision making

Core Features

• Tick Data Integration: Direct connection to TimescaleDB for efficient retrieval of historical and streaming tick data
• Multi-Strategy Evaluation: Concurrent simulation of multiple trading robots with correlation analysis
• MFE-Based Exit Logic: Dynamic trade exits based on portfolio-level Maximum Favorable Excursion
• Real-Time Analytics: Continuous calculation of key risk metrics and performance indicators
• Modular Architecture: Easily extendable with custom strategies and analytics

System Architecture

The SDK consists of four primary components:

1. Data Layer: Efficient tick data retrieval from TimescaleDB
2. Strategy Manager: Multi-strategy simulation and position tracking
3. Portfolio Evaluator: Collective performance analysis and MFE-based exit rules
4. Execution Module: Trade simulation and performance reporting

Installation

# Clone the repository
git clone https://github.com/your-username/quantitative-trading-portfolio-sdk.git

# Create a Python virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

Usage

from data_science import TickDataProvider, StrategyManager, PortfolioEvaluator, ExecutionSimulator

# Initialize components
tick_provider = TickDataProvider("postgresql://user:password@localhost:5432/tickdb")
strategy_manager = StrategyManager()
portfolio_evaluator = PortfolioEvaluator(strategy_manager, {"mfe_exit_threshold": 0.3})
execution_simulator = ExecutionSimulator(strategy_manager, portfolio_evaluator)

# Register trading strategies
strategy_manager.register_strategy("rsi_strategy", rsi_strategy_function)
strategy_manager.register_strategy("macd_strategy", macd_strategy_function)

# Run backtest with historical data
historical_ticks = tick_provider.get_tick_range("PETR4", "2023-01-01", "2023-01-31")
for tick in historical_ticks.to_dict('records'):
    metrics = execution_simulator.process_tick_data(tick)