Skip to main content

General Settings

These settings form the baseline configuration for the Lorentzian Classification model. They control aspects like data sources, how far into the past the model looks, how it calculates distances, how trades are displayed and timed, and the complexity of the training logic.

Interactive Settings Explorer​

Below is an interactive exploration of the General Settings interface. Click any setting to instantly view its detailed documentation, default value, and purpose - a quick alternative to scrolling through the full reference below:

Lorentzian Classification [2.06 - Beta]

🔧 GENERAL SETTINGS

Setting Help

Select a setting to view its help information

Detailed Settings Reference​

Core Configuration Settings​

Reduce Price-Time Warping

Default: false
Purpose

Minimizes distortions in price-time relationships during pattern matching.

Impact

Affects how the model identifies and weighs historical patterns.

Recommendations
  • Enable for markets with high volatility
  • Disable for range-bound markets
  • Test both settings in your specific market context

Source

Default: close
Purpose

Determines which price point of each bar is used for calculations (close, open, high, or low).

Impact

The choice of price source can significantly affect how the indicator interprets market movements and generates signals.

Recommendations
  • Use 'close' for most reliable signals
  • Consider 'high/low' for range-based strategies
  • Match source to your trading timeframe
Best Practice

Stick with the default close source unless you have a specific reason to change it. This ensures optimal compatibility with the model's core algorithms.

Neighbor Configuration​

Neighbors Count

Default: 8
Purpose

Sets the number of nearest neighbors used in the classification algorithm. More neighbors can provide smoother signals but may reduce sensitivity.

Impact

Higher values create more stable but slower signals, while lower values increase sensitivity but may generate more noise.

Recommendations
  • Start with the default of 8 neighbors
  • Increase for less volatile markets
  • Decrease for more responsive signals

Max Bars Back

Default: 2000
Purpose

Defines the maximum number of historical bars used for calculations. Affects both accuracy and performance.

Impact

Larger values provide more historical context but require more computational resources.

Recommendations
  • Use at least 1000 bars for reliable signals
  • Adjust based on your trading timeframe
  • Consider your platform's performance capabilities

Exit Settings​

Show Default Exits

Default: true
Purpose

Displays the indicator's built-in exit signals based on kernel regression analysis.

Impact

Helps identify potential exit points for trades, improving risk management.

Recommendations
  • Keep enabled for complete trading signals
  • Use in conjunction with your own exit rules
  • Compare with dynamic exits for confirmation

Use Dynamic Exits

Default: false
Purpose

Calculates exit points dynamically based on market volatility and trend strength.

Impact

Provides more adaptive exit signals that adjust to changing market conditions.

Recommendations
  • Enable for volatile markets
  • Combine with fixed exits for confirmation
  • Monitor effectiveness in your timeframe

Advanced Configuration​

Reset Factor

Default: 0.75
Purpose

Controls how aggressively the algorithm resets its internal state when market conditions change significantly.

Impact

Higher values make the indicator more responsive to changes but may increase false signals.

Recommendations
  • Lower values for trending markets
  • Higher values for ranging markets
  • Adjust based on false signal frequency
Technical Detail

Lower reset factors create more selective neighbor choices, while higher values are more lenient in maintaining historical relationships.

Use Downsampling

Default: false
Purpose

Reduces data resolution to improve performance while maintaining signal quality.

Impact

Can significantly improve performance on larger datasets with minimal impact on accuracy.

Recommendations
  • Enable for very large datasets
  • Use when experiencing performance issues
  • Consider disabling for precise entry/exit timing

Performance Monitoring​

Show Trade Stats

Default: true
Purpose

Displays real-time statistics about the indicator's performance and trade signals.

Impact

Provides immediate feedback on the indicator's effectiveness and helps in strategy refinement.

Recommendations
  • Keep enabled during strategy development
  • Use to validate signal quality
  • Monitor for strategy optimization

Use Worst Case Estimates

Default: false
Purpose

Calculates signals using conservative estimates to reduce false positives.

Impact

Generates fewer but potentially more reliable signals.

Recommendations
  • Enable for risk-averse strategies
  • Use in volatile market conditions
  • Combine with other confirmation tools

Pattern Recognition​

Use Remote Fractals

Default: true
Purpose

Incorporates fractal pattern analysis from higher timeframes for signal generation.

Impact

Improves signal quality by considering larger market structures.

Recommendations
  • Keep enabled for trend following
  • Especially useful in higher timeframes
  • Monitor impact on signal timing

Default Configuration Rationale​

The default settings are carefully chosen to provide:

  • Simplified initial setup
  • Clear base exit logic
  • Rich historical context
  • Balanced performance metrics
  • Optimistic preview mode
Why These Defaults?

Default values prioritize ease of understanding while maintaining the indicator's core strengths. They provide a solid foundation for both learning and customization.

Summary​

These general settings provide the foundation for Lorentzian Classification's behavior. While the defaults offer a balanced starting point, understanding each setting's purpose allows for precise tuning to match your trading style and market conditions.

Try experimenting with the interactive demo above to see how different settings affect the configuration. Each setting has detailed documentation that appears when you hover over it in the interface.

Example Configurations​

Quick & Simple Setup:​

  • Reduce Price-Time Warping: False - Reasoning: Maintains the indicator's core sensitivity to market volatility, allowing it to react quickly to sudden price changes.
  • Source: close - Reasoning: Uses the standard closing price for broad compatibility, ensuring the indicator works consistently across different assets and exchanges.
  • Neighbors Count: 8 - Reasoning: Provides a balanced number of neighbors for classification, offering a good trade-off between responsiveness and noise reduction.
  • Max Bars Back: 2000 - Reasoning: Offers a reasonable historical context without excessive computation, allowing the indicator to analyze a sufficient amount of past data without slowing down performance.
  • Show Exits: False - Reasoning: Simplifies the chart by hiding exit signals, reducing visual clutter and focusing on entry points.
  • Use Dynamic Exits: False - Reasoning: Uses a fixed exit strategy for simplicity, making it easier to understand and manage trades.
  • Use Remote Fractals: True - Reasoning: Leverages a broader historical context for pattern recognition, potentially identifying more reliable signals.
  • Show Trade Stats: True - Reasoning: Provides basic performance feedback, allowing users to quickly assess the effectiveness of the configuration.
  • Use Worst Case Estimates: False - Reasoning: Shows a more optimistic view of potential performance, encouraging experimentation and exploration. This configuration is designed for New to ML/Trading Beginners seeking an easy starting point. It uses default settings for most parameters, providing a balanced configuration for general use and easy exploration of the indicator's basic functionality. Remote Fractals are enabled to provide a broader historical context, while keeping other settings at their defaults allows for a less overwhelming initial experience.

Robust Trend Confirmation:​

  • Reduce Price-Time Warping: True - Reasoning: Reduces sensitivity to short-term volatility, emphasizing longer trends and providing more stable signals.
  • Source: close - Reasoning: Uses the standard closing price for broad compatibility, ensuring consistent results across different assets and exchanges.
  • Neighbors Count: 12 - Reasoning: Increases the number of neighbors for more robust trend confirmation, reducing the impact of outliers and improving signal reliability.
  • Max Bars Back: 3000 - Reasoning: Provides a richer historical context for trend analysis, allowing the model to learn from a wider range of market conditions and identify more reliable trends.
  • Show Exits: True - Reasoning: Displays exit signals for clear trade management, enabling users to visualize potential exit points and manage risk effectively.
  • Use Dynamic Exits: True - Reasoning: Adapts exit points to changing market conditions, potentially improving profitability and reducing losses.
  • Use Remote Fractals: False - Reasoning: Focuses on recent price action for trend detection, reducing the influence of distant historical patterns that may no longer be relevant.
  • Show Trade Stats: True - Reasoning: Provides performance feedback to monitor the effectiveness of the configuration and track key metrics like win rate and drawdown.
  • Use Worst Case Estimates: True - Reasoning: Provides a conservative view of potential performance, aligning with risk-averse strategies and ensuring a realistic assessment of potential losses. This configuration is designed for Fixed Income/Conservative Traders and Backtest Optimizers prioritizing robust trend confirmation. It reduces price-time warping for more Euclidean character, increases Neighbors Count and Max Bars Back for robust trend confirmation, and enables Dynamic Exits and Worst Case Estimates for conservative backtesting. The focus is on minimizing risk and maximizing the reliability of trend-following signals, making it suitable for long-term investors and those with a low-risk tolerance.

Optimized for Scalping:​

  • Reduce Price-Time Warping: False - Reasoning: Maintains sensitivity to short-term price fluctuations, crucial for identifying quick scalping opportunities.
  • Source: close - Reasoning: Uses the standard closing price for broad compatibility, ensuring consistent results across different assets and exchanges.
  • Neighbors Count: 5 - Reasoning: Reduces the number of neighbors for faster signal generation, essential for quick entries and exits in fast-moving markets.
  • Max Bars Back: 1000 - Reasoning: Minimizes computational load and emphasizes recent price action, improving responsiveness and reducing lag.
  • Show Exits: False - Reasoning: Simplifies the chart for quick decision-making, reducing visual clutter and allowing for faster reactions to market changes.
  • Use Dynamic Exits: True - Reasoning: Adapts exit points to fast-moving market conditions, allowing for quicker profit-taking or loss-cutting and maximizing efficiency.
  • Use Remote Fractals: True - Reasoning: Leverages a broader historical context for pattern recognition, potentially identifying subtle scalping opportunities that might be missed with a shorter lookback period.
  • Show Trade Stats: True - Reasoning: Provides performance feedback to monitor the effectiveness of the configuration and track key metrics like win rate and profit factor.
  • Use Worst Case Estimates: False - Reasoning: Prioritizes speed over conservative backtesting, reflecting the aggressive nature of scalping and focusing on potential gains rather than worst-case scenarios. This configuration is tailored for Performance Tweakers and Reactive Intraday Traders focused on low-latency scalping. It reduces Neighbors Count and Max Bars Back to minimize computational load and increase responsiveness for scalping. Dynamic Exits are enabled for quicker exits, while Worst Case Estimates are disabled to prioritize speed over conservative backtesting. The goal is to generate frequent, timely signals for quick profit opportunities, even if it means accepting a higher risk of false signals.