trading_bot_v4/docs/analysis/README.md

Performance Analysis & Optimization Studies

Data-driven insights for trading system optimization.

This directory contains performance analyses, blocked signals tracking, profit projections, and diagnostic results. All files focus on measuring system performance and identifying optimization opportunities.

---

📊 Key Documents

Performance Reviews

• PROFIT_PROJECTION_NOV24_2025.md - Monthly profit projections and compound growth analysis
• QUALITY_90_TP1_ONLY_ANALYSIS.md - Analysis of quality 90+ signals with TP1-only exits
• SIGNAL_QUALITY_TEST_RESULTS.md - Signal quality scoring validation results

Blocked Signals Analysis

• BLOCKED_SIGNALS_TRACKING.md - System for tracking signals blocked by quality filters
• BLOCKED_SIGNAL_ANALYSIS_DEC2.md - December 2 analysis of blocked signals
• BLOCKED_SIGNAL_CORRECTED_ANALYSIS_DEC2.md - Corrected analysis after data quality fixes

Adaptive Leverage

• LONG_ADAPTIVE_LEVERAGE_VERIFICATION.md - Verification of LONG direction adaptive leverage
• SHORT_ADAPTIVE_LEVERAGE_IMPLEMENTED.md - SHORT direction adaptive leverage implementation

System Health

• DIAGNOSTIC_RESULTS_SUMMARY.md - Comprehensive diagnostic test results
• SAFETY_ANALYSIS.md - Risk management and safety parameter analysis
• RECOVERY_PLAN.md - System recovery procedures after incidents
• ANALYTICS_STATUS_AND_NEXT_STEPS.md - Analytics dashboard status and roadmap

Strategic Planning

• CLEANUP_PLAN.md - Documentation reorganization strategy (precursor to this structure)

---

📈 Analysis Workflow

1. Data Collection:

• Trades executed with quality scores, ATR, ADX, RSI metrics
• Blocked signals saved to BlockedSignal table for comparison
• Position Manager tracks MAE (Max Adverse Excursion) and MFE (Max Favorable Excursion)

2. Analysis Queries:

-- Win rate by quality tier
SELECT 
  CASE 
    WHEN signalQualityScore >= 95 THEN '95-100'
    WHEN signalQualityScore >= 90 THEN '90-94'
    WHEN signalQualityScore >= 85 THEN '85-89'
    ELSE '60-84'
  END as tier,
  COUNT(*) as trades,
  ROUND(100.0 * SUM(CASE WHEN realizedPnL > 0 THEN 1 ELSE 0 END) / COUNT(*), 1) as win_rate,
  ROUND(SUM(realizedPnL), 2) as total_pnl
FROM "Trade"
WHERE exitReason IS NOT NULL
GROUP BY tier
ORDER BY MIN(signalQualityScore) DESC;

3. Optimization:

• Compare blocked vs executed signal performance
• Identify quality threshold sweet spots
• Validate parameter changes with 50-100 trade samples
• Document findings in this directory

---

🎯 Current Focus (Dec 2025)

Active Analyses:

• Quality threshold optimization (LONG 90, SHORT 95 under evaluation)
• Adaptive leverage performance tracking (10x vs 5x tiers)
• Blocked signal win rate comparison (would-be winners vs actual blocks)
• Runner system effectiveness (40% runner with ATR trailing stop)

Upcoming:

• Multi-timeframe comparison (5min vs 15min vs 1H)
• MA crossover pattern validation (ADX weak→strong hypothesis)
• Smart Entry Validation Queue effectiveness tracking

---

📝 Adding New Analyses

Template Structure:

# [Analysis Title]

**Date:** [YYYY-MM-DD]
**Data Range:** [Date range or trade count]
**Objective:** [What you're analyzing]

## Executive Summary
[2-3 sentence key findings]

## Methodology
[How data was collected, SQL queries, filters applied]

## Results
[Tables, charts, statistics]

## Insights
[What the data means, patterns discovered]

## Recommendations
[Specific actions, parameter changes, threshold adjustments]

## Verification Required
[How to test recommendations, sample sizes needed]

Naming Convention:

• [TOPIC]_ANALYSIS_[DATE].md for dated analyses
• [FEATURE]_VERIFICATION.md for feature validation
• [METRIC]_TRACKING.md for ongoing measurement

---

🔍 Finding Specific Analyses

By Topic:

• Quality scores → SIGNAL_QUALITY_*, QUALITY_90_*
• Leverage → *_ADAPTIVE_LEVERAGE_*
• Blocked signals → BLOCKED_SIGNAL_*
• System health → DIAGNOSTIC_*, SAFETY_*, RECOVERY_*

By Date:

• All files include creation dates in content
• Most recent: December 2025 blocked signal analyses
• Historical: CLEANUP_PLAN.md (documentation strategy)

---

⚠️ Important Notes

Data Integrity:

• Always filter out signalSource='manual' for indicator analysis
• Use WHERE timeframe='5' for production signal analysis
• Exclude blockReason='DATA_COLLECTION_ONLY' for execution analysis

Sample Sizes:

• Minimum 50 trades for statistical validity
• Minimum 20 blocked signals for threshold analysis
• 100+ trades preferred for major parameter changes

Real Money System:

• Every analysis affects financial outcomes
• Always include "Expected Impact" section (dollar amounts)
• Document both upside potential and downside risk
• Require user approval for threshold changes >10 points

---

See ../README.md for overall documentation structure.