SELF-LEARNING PREDICTION MACHINE

Workshop Tutorial: Bitcoin Prediction, Backtesting, Correlation Analysis & Causation Studies

Read the Paper on blog.macropulze.com

Historical parameter optimization and performance analysis

Backtest Configuration

How Backtesting Works

Backtesting tests your prediction model on historical data to evaluate performance. The system:

  • Tests different parameter combinations
  • Measures accuracy, profitability, and risk
  • Identifies optimal settings for live trading
  • Provides trading suitability assessment

⚠️ Daily Limit: Only one backtest per day is allowed to conserve system resources. Cached results will be shown for subsequent attempts on the same day.

Status & Progress

Configure dates and parameters above, then click "Run Backtest" to start analysis.

Quick Test: Tests ~100 parameter combinations
Grid Search: Tests ~1000 parameter combinations
Random Search: Tests ~500 random combinations

Backtest Results & Analysis

Understanding Your Results

Results show how your model would have performed on historical data. Key metrics include:

  • Accuracy: % of correct price direction predictions (>55% is good)
  • Profit Factor: Total gains ÷ total losses (>1.5 is good)
  • Sharpe Ratio: Risk-adjusted returns (>1.0 is good)
  • Max Drawdown: Largest loss from peak (<20% is acceptable)

Run a backtest to see detailed performance analysis and trading suitability assessment

Backtest History

Loading history...

Select backtests from history to compare performance metrics