Navigation

MangoLabsDOCS
Launch

Managing Features

Learn how to effectively manage your feature library in MangoLabs. This guide covers viewing, editing, deleting, visualizing, and organizing your technical indicators for optimal strategy development and performance.

What You'll Learn

  • Navigating and viewing your feature library
  • Editing existing features safely
  • Deleting features and understanding dependencies
  • Visualizing features on charts
  • Organization strategies and best practices

Viewing Your Feature Library

The feature library is your central hub for managing all technical indicators:

Accessing Your Features

  1. Click "Features" in the left sidebar to open your feature library
  2. View all your created features displayed as cards
  3. Each card shows: Feature name, indicator type, parameters, and usage count
  4. Use the search bar to quickly find specific features
  5. Filter by indicator type (RSI, SMA, MACD, etc.) to narrow results

Feature Card Information

Name

The unique identifier (e.g., RSI_14)

Type

Indicator type (RSI, SMA, Bollinger Bands, etc.)

Parameters

Configuration values (period, source, etc.)

Usage

Number of strategies using this feature

Quick Actions

View on Chart

Visualize the indicator on historical data

Edit

Modify feature parameters (with warnings)

Delete

Remove feature from library (check dependencies first)

Editing Features

You can edit existing features, but be aware of the impact on strategies using them:

How to Edit a Feature

  1. Navigate to the feature card in your library
  2. Click the "Edit" button (pencil icon)
  3. Modify the parameters you want to change
  4. Review the warning about strategies that use this feature
  5. Click "Save Changes" to apply

When Editing is Safe

Editing features is safe in these scenarios:

  • Name Changes: Updating the feature name doesn't affect calculations, only identification
  • Description Updates: Adding or modifying the description is always safe
  • Unused Features: If no strategies use the feature (usage count = 0), edit freely

When Editing Breaks Strategies

Changing these parameters will affect all strategies using the feature:

  • Period/Length: Changing RSI from 14 to 21 will produce completely different values
  • Source: Switching from "close" to "high" changes what price data is analyzed
  • Method: Changing SMA to EMA fundamentally alters the calculation

Recommendation: Instead of editing, create a new feature with the desired parameters (e.g., RSI_21 instead of modifying RSI_14). This preserves existing strategies.

Deleting Features

Removing features from your library requires understanding dependencies and impact:

Deletion Process

  1. Click the "Delete" button (trash icon) on the feature card
  2. Review the dependency warning showing which strategies use this feature
  3. If dependencies exist, you'll see a list of affected strategies
  4. Confirm deletion only if you understand the impact
  5. The feature will be permanently removed from your library

Safe to Delete

  • Feature has 0 strategies using it
  • Feature was created for testing purposes
  • Feature is a duplicate of another feature
  • Feature uses incorrect parameters

Dangerous to Delete

  • Feature is used by active strategies
  • Feature is used in paper trading
  • Feature has extensive backtest history
  • Multiple strategies depend on it

What Happens When You Delete a Used Feature?

Strategies Break: Any strategy using the deleted feature will fail to execute. The strategy canvas will show missing feature errors.

Backtests Become Invalid: Historical backtest results remain visible, but you can't re-run them without the feature.

Paper Trading Stops: If a deployed strategy uses the deleted feature, it will stop generating signals and may error out.

Recovery: The only way to fix broken strategies is to recreate the feature with identical parameters or remove the feature from all affected strategies.

Visualizing Features on Charts

Chart visualization helps you understand how your features behave with real market data:

Using Chart Preview

  1. Click "View on Chart" button on any feature card
  2. Select a trading pair (e.g., BTCUSDT, ETHUSDT)
  3. Choose a timeframe (1m, 5m, 15m, 1h, 4h, 1d)
  4. The chart displays historical price data with your indicator overlaid
  5. Interact with the chart: zoom, pan, hover for values
  6. Toggle indicator visibility or adjust chart display settings

What to Look For

  • Value Ranges: RSI should be 0-100, moving averages should follow price
  • Smoothness: Indicators should smooth price action appropriately
  • Lag: Notice how much the indicator lags behind price changes
  • Signals: Identify where buy/sell signals would occur

Chart Analysis Tips

  • Multiple Timeframes: Check indicator on different timeframes
  • Different Symbols: Test how it behaves across various assets
  • Historical Events: See how indicator performed during crashes/rallies
  • Compare Features: View multiple features together for strategy ideas

Pro Tip: Use chart visualization to experiment with different parameter values before committing. Create test features with varying periods (RSI_9, RSI_14, RSI_21) and compare their signals on the same chart to find what works best for your strategy.

Organization Best Practices

As your feature library grows, organization becomes crucial for efficiency:

Naming Strategy

Use Descriptive, Consistent Names

Good Examples

  • RSI_14_Close
  • SMA_50_Daily
  • BB_20_2_StdDev
  • MACD_12_26_9

Bad Examples

  • my_indicator_v3
  • test_feature_2
  • rsi
  • moving_avg

Format Template: INDICATOR_PARAM1_PARAM2_DESCRIPTION
This makes it easy to identify features at a glance and understand their configuration without opening details.

Feature Grouping Strategies

By Indicator Type

  • All RSI variants together
  • All moving averages together
  • All volatility indicators together

By Strategy Use

  • Prefix with strategy name (e.g., SCALP_RSI_9)
  • Group features used in same strategy
  • Mark shared features clearly

By Timeframe

  • Include timeframe in name
  • Separate 1h features from 4h features
  • Clarify which timeframe they're optimized for

By Purpose

  • Entry signals vs exit signals
  • Trend confirmation vs timing
  • Risk management vs signal generation

Maintenance Routine

Weekly Review

Check for unused features (usage count = 0) and consider deleting test features that are no longer needed.

Monthly Cleanup

Identify duplicate features with slightly different parameters and consolidate where possible.

Documentation Updates

Ensure all features have clear descriptions explaining their purpose and which strategies use them.

Performance and Cache Management

Understanding how features are cached helps you optimize system performance:

How Feature Caching Works

Initial Calculation

When a feature is first used in a strategy, it's calculated for all historical data and stored in the database.

Reuse Benefit

If another strategy uses the same feature on the same symbol/timeframe, values are retrieved from cache instantly—no recalculation needed.

Real-Time Updates

In paper/live trading, features update incrementally as new candles close, maintaining cache efficiency.

Good for Performance

  • Reusing features across multiple strategies
  • Creating features once with standard parameters
  • Deleting unused features to reduce storage
  • Using cached data for backtests

Bad for Performance

  • Creating slight variations (RSI_14, RSI_14_v2, RSI_14_final)
  • Constantly editing feature parameters (invalidates cache)
  • Keeping hundreds of unused test features
  • Creating features with extreme periods (e.g., SMA_5000)

Performance Optimization Tips

  • Standardize Parameters: Use common industry standards (RSI-14, SMA-50, BB-20) so features are reusable across strategies.
  • Check Before Creating: Always search your library before creating a new feature—someone might have already created what you need.
  • Clean Up Regularly: Delete test features and unused variations to keep the library lean and performant.
  • Avoid Over-Optimization: Creating 20 variations of RSI with slightly different periods doesn't significantly improve strategies but hurts performance.

Common Issues and Solutions

Issue: Feature shows NaN values on chart

Cause: Insufficient historical data for the indicator period. For example, RSI-14 needs at least 14 candles.

Solution: Ensure you have enough historical data loaded. Check if the period parameter is too large for the available data range.

Issue: Strategy won't run after editing feature

Cause: Editing feature parameters invalidated cached data, causing strategy to fail.

Solution: Recreate the feature with original parameters or update the strategy to use a new feature. Always prefer creating new features over editing existing ones.

Issue: Can't delete feature (dependency error)

Cause: One or more strategies are currently using this feature.

Solution: Review the dependency list. Either remove the feature from all dependent strategies first, or keep the feature if it's actively used.

Issue: Feature library loading slowly

Cause: Too many features in the library (hundreds of test features, duplicates).

Solution: Perform library cleanup. Delete unused test features, consolidate duplicates, and organize with clear naming to reduce clutter.

What's Next?

Back to Feature Fundamentals

Review what features are and how they work in the MangoLabs system.

Create More Features

Learn how to create new features with proper parameters and naming conventions.

Build a Strategy

Use your well-organized feature library to build and test trading strategies.