Combined Analytics Guide

Overview

Combined Analytics allows you to analyze how multiple components interact with each other. Instead of viewing component statistics in isolation, you can discover which combinations of selections lead to the best (or worst) results.

Example: Rather than just knowing that "Long" trades have a 60% winrate, Combined Analytics can show you that "Long + High Confidence + Setup A" has a 75% winrate, while "Long + Low Confidence + Setup B" only has 45%.

Accessing Combined Analytics

• Navigate to the Statistics page
• In the Filter Panel (right side), click the Combination Analysis button
• The Combined Analytics overlay will open

Combination Types

Inter-Component Combinations (Current)

Analyze how selections across different components correlate with performance.

Example: Button component "Direction" (Long/Short) combined with Dropdown "Setup Type" (Breakout/Pullback/Reversal)

Intra-Component Combinations (Coming Soon)

Analyze how multiple selections within the same component correlate with performance.

Example: When using a multi-select Checkbox component for "Trade Criteria", see performance when both "Criterion A" AND "Criterion B" are checked together.

Selecting Components

Component Selection

1. Click the component selector dropdown
2. Choose 2 to 4 components to analyze
3. Selected components appear as chips below the selector
4. Click the X on a chip to remove it
5. Click Analyze Combinations when ready

Recommendations

• 2 components: Easiest to interpret, best for beginners
• 3 components: Good balance of detail and clarity
• 4 components: Maximum depth, but can be complex to interpret

Understanding the Results

Performance Matrix (Blue Heatmap)

When your selected components don't include multiple button components with win/loss states:

• Color: Blue gradient (light blue = low winrate, dark blue = high winrate)
• Shows: Winrate percentage for each combination
• Based on: Global Trade Result or Net PnL component

Usage Matrix (Purple Heatmap)

When you select 2 or more button components with win/loss states:

• Color: Purple gradient (light purple = rare, dark purple = frequent)
• Shows: Usage frequency (how often state combinations occur)
• Why? With multiple button components having states, it's impossible to determine which button's state should define the "win" or "loss" - so we show usage instead

Example Scenario:

If you analyze two button components:

• Component 1: "First Entry" with states: Win, Loss, Break Even
• Component 2: "Reverse Entry" with states: Win, Loss, Break Even

Both can have different states in the same trade (First Entry might be Win, while Reverse Entry is Loss). Since we can't determine which state represents the "overall" win/loss, the system switches to showing usage frequency - how often each combination occurs in your data.

You can then cross-reference with your global results to calculate winrates manually if needed.

Visualization Tabs

Combined Analytics provides 4 ways to view your combination data:

1. Heatmap (Recommended for Overview)

Best for: Getting a quick visual overview of all combinations

• Visual color-coding makes patterns immediately obvious
• Works great for 2-component analysis (2D grid)
• For 3+ components, select which two to display on X/Y axes
• Hover over cells for detailed stats

2. Data Table (Best for Top Performers)

Best for: Finding your highest/lowest performing combinations

• Sortable by winrate or trade count via dropdown menu
• Shows all combinations in one scrollable list
• Searchable by combination name
• Perfect for identifying your "best" and "worst" setups

3. Top Performers (Bar Chart)

Best for: Visual comparison of best combinations

• Bar chart showing top 10, 20, or 50 combinations
• Sorted by winrate (highest first)
• Bar length represents winrate percentage
• Hover for detailed stats
• Good for presentations or quick comparisons

4. Summary (Overview Stats)

Best for: High-level summary of the analysis

• Total combinations found
• Best combination (highest winrate with minimum sample size)
• Worst combination (lowest winrate)
• Average winrate across all combinations
• Total trades analyzed

Interpreting Results

How Winrates are Calculated

For Button Components with Win/Loss States

If the button component itself has win/loss/break even states configured:

• The button's state directly determines the outcome
• Does NOT rely on global Trade Result or Net PnL
• Each button state is independent

Example: Button "Direction" with states: Long(Win), Short(Loss) - The "(Win)" and "(Loss)" states are used directly.

For Other Components (Checkboxes, Sliders, Dropdowns, etc.)

These components 100% rely on your global Trade Result or Net PnL component:

• When you enter a trade with global result = Win and Net PnL positive, the entry counts as a win
• All component selections in that trade are associated with the win
• The system calculates: "When X was selected, how often did the trade result in a win?"

Real-World Examples

Example 1: Checkbox Performance

Scenario: You have a Checkbox component with 4 criteria: A, B, C, D

You enter a trade where:

• Criteria A, B, C are checked (D is not)
• Global Trade Result = Win
• Net PnL = +$150

Result: This trade counts as a WIN for:

• Combination "A + B" (2 of 3)
• Combination "A + C" (2 of 3)
• Combination "B + C" (2 of 3)
• Combination "A + B + C" (all 3)

The system tracks: "When A and B were both checked, what was the winrate?"

Example 2: Mixed Components

Scenario: Analyzing Button "Direction" (Long/Short) + Dropdown "Session" (London/NY/Asian)

You enter a trade:

• Direction = Long
• Session = NY
• Global Trade Result = Win

Result: This adds one WIN to the combination "Long + NY"

Combined Analytics will show: "Long + NY has a 68% winrate over 42 trades"

Example 3: Button States (Multiple Buttons)

Scenario: Two button components with win/loss states

• Button 1 "First Entry": Win state
• Button 2 "Scale-in Entry": Loss state

Result: Since both buttons have states, the system shows purple usage heatmap instead of winrate. You'll see "First(Win) + Scale-in(Loss) occurred 12 times (8% of total)"

You can manually calculate: "Of those 12 times, check global results to see overall trade outcome."

Sample Size Considerations

• Minimum 5 trades recommended before trusting a combination's stats
• Small sample sizes can be misleading (3/3 wins = 100%, but not statistically significant)
• Focus on combinations with at least 10-20 trades for reliable patterns
• The "Total" column shows how many trades each combination represents

Exporting Data

• Click Export CSV to download combination data as a spreadsheet
• Click Export JSON to download raw data for custom analysis
• Exports include all combinations, winrates, and trade counts

Best Practices

• Start with 2 components - Easiest to understand and interpret
• Use components you actually vary - If you always trade "Long", don't analyze Direction
• Ensure sufficient data - Need at least 50-100 total trades for meaningful analysis
• Focus on patterns - Look for combinations that significantly outperform or underperform
• Validate findings - If a combination shows 90% winrate, review the actual trades to understand why
• Consider context - A winning combination in one market condition may not work in another
• Use filters wisely - Apply date range filters to analyze specific periods

Common Questions

Why is the heatmap purple instead of blue?

Answer: Purple indicates Usage Mode, which happens when you've selected 2 or more button components that have win/loss states. Since we can't determine which button's state should define the overall outcome, we show how often combinations occur instead of winrate.

Can I analyze more than 4 components?

Answer: The system supports up to 4 components. More than 4 creates exponentially more combinations (making the data spread too thin) and becomes very difficult to interpret meaningfully.

What if a combination has no data?

Answer: The heatmap shows a dash for combinations that never occurred in your data. This is normal - not all possible combinations will exist in your trading history.

Why don't I see a heatmap for 3+ components?

Answer: For 3+ components, you need to select which 2 components to display on the X and Y axes. Use the axis selectors to choose, and the heatmap will display that 2D slice of your data. The Table view shows all combinations regardless of component count.

How is "Best Combination" determined?

Answer: Best combination is the one with the highest winrate that has at least 5 trades (minimum sample size). This prevents combinations with only 1-2 trades from being labeled "best" due to small sample size luck.

Can I see individual trades for a combination?

Answer: Not directly in Combined Analytics. However, you can note the combination (e.g., "Long + NY + Breakout") and then use the Archive page to filter and review those specific trades.

Troubleshooting

No Combinations Found

Problem: Analysis returns 0 combinations.

Solutions:

• Verify the selected components exist in your chosen profile and layout
• Check that you have trades entered with those components
• Ensure your date range includes trades with the selected components
• Try a broader date range or different components

All Combinations Show Low Trade Counts

Problem: Every combination only has 1-3 trades.

Solutions:

• You may have selected too many components (4 creates many combinations)
• Try analyzing fewer components (2-3 instead of 4)
• Ensure you have enough total trades entered (need 50+ for meaningful multi-component analysis)

Unexpected Usage Matrix

Problem: Getting purple usage matrix when you expected blue winrate.

Solutions:

• Check if you've selected 2+ button components that have win/loss states
• If you want winrate analysis, select only 1 button component with states, or select other component types
• If you intentionally selected multiple button components, usage matrix is correct behavior

Related Topics

• Statistics Guide
• Layout Editor Guide
• Getting Started