docs: Add Common Pitfall #72 - MFE data unit mismatch

CRITICAL LESSON LEARNED (Dec 5, 2025):
Document the data analysis disaster caused by MFE/MAE stored in mixed units.

What Happened:
- Analyzed blocked vs executed signals to improve win rate
- SQL showed executed signals: 20.15% avg MFE (appeared excellent)
- Implemented "optimizations" based on this data:
  * Tighter ATR multipliers (2.0→1.5, 4.0→3.0)
  * Higher TP1 close (60%→75%)
  * Increased leverage (1×→5×)
- User questioned: Why doesn't TP1 hit if it's 20% MFE?
- Investigation: Only 2/11 trades reached TP1 price target
- Root cause: Old records stored MFE in DOLLARS, new in PERCENTAGES
- TRUE MFE: 0.76% (long), 1.20% (short) - NOT 20%!
- 26× inflation due to unit mismatch

Why This Matters:
- This is a REAL MONEY system - wrong analysis = wrong trades = losses
- MFE/MAE used for critical decisions (exit timing, quality validation)
- Agent made "data-driven" optimizations on 26× inflated data
- All changes had to be reverted (commits a67a338, f65aae5)

MANDATORY SQL Pattern:
- ALWAYS filter by createdAt >= '2025-11-23' for MFE/MAE queries
- OR recalculate from prices (maxFavorablePrice - entryPrice)
- NEVER trust raw AVG(maxFavorableExcursion) without date filter

Prevention:
- Verify stored vs calculated values before ANY MFE/MAE analysis
- Check sample of recent vs old records to detect unit changes
- Document data format changes in Common Pitfalls immediately

Related:
- Common Pitfall #54: Original MFE/MAE units bug (Nov 23, 2025)
- Revert commit: a15f17f
- Incorrect optimization: a67a338, f65aae5

.github/copilot-instructions.md

@@ -2817,7 +2817,75 @@ This section contains the **TOP 10 MOST CRITICAL** pitfalls that every AI agent
 
 📚 **Full Documentation:** `docs/COMMON_PITFALLS.md` (72 pitfalls with code examples, git commits, deployment dates)
 
-72. **CRITICAL SECURITY: .env file tracked in git (CRITICAL - Fixed Dec 5, 2025 - PR #3):**
+72. **CRITICAL: MFE Data Unit Mismatch - ALWAYS Filter by Date (CRITICAL - Dec 5, 2025):**
+    - **Symptom:** SQL analysis shows "20%+ average MFE" but TP1 (0.6% target) never hits
+    - **Root Cause:** Old Trade records stored MFE/MAE in DOLLARS, new records store PERCENTAGES
+    - **Data Corruption Examples:**
+      * Entry $126.51, Peak $128.21 = 1.35% actual move
+      * But stored as maxFavorableExcursion = 90.73 (dollars, not percent)
+      * SQL AVG() returns meaningless mix: (1.35 + 90.73 + 0.85 + 87.22) / 4 = 45.04
+    - **Incident (Dec 5, 2025):**
+      * Agent analyzed blocked vs executed signals
+      * SQL showed executed signals: 20.15% avg MFE (appeared AMAZING)
+      * Implemented "optimizations": tighter targets, higher TP1 close, 5× leverage
+      * User questioned: "tp1 barely hits that has nothing to do with our software monitoring does it?"
+      * Investigation revealed: Only 2/11 trades reached TP1 price
+      * TRUE MFE after filtering: 0.76% (long), 1.20% (short) - NOT 20%!
+      * 26× inflation due to unit mismatch in old data
+    - **MANDATORY SQL Pattern:**
+      ```sql
+      -- WRONG: Includes corrupted old data
+      SELECT AVG("maxFavorableExcursion") FROM "Trade" 
+      WHERE "signalQualityScore" >= 90;
+      
+      -- CORRECT: Filter to after Nov 23, 2025 fix
+      SELECT AVG("maxFavorableExcursion") FROM "Trade" 
+      WHERE "signalQualityScore" >= 90 
+        AND "createdAt" >= '2025-11-23';  -- After MFE fix
+      
+      -- OR: Recalculate from prices (always correct)
+      SELECT AVG(
+        CASE 
+          WHEN direction = 'long' THEN 
+            (("maxFavorablePrice" - "entryPrice") / "entryPrice") * 100
+          ELSE 
+            (("entryPrice" - "maxFavorablePrice") / "entryPrice") * 100
+        END
+      ) FROM "Trade" WHERE "signalQualityScore" >= 90;
+      ```
+    - **Why This Matters:**
+      * **This is a REAL MONEY system** - wrong analysis = wrong trades = financial losses
+      * MFE/MAE used for exit timing optimization, trade analysis, quality validation
+      * Agent made "data-driven" decisions based on 26× inflated numbers
+      * Optimizations REVERTED via commits f65aae5 and a67a338 (Dec 5, 2025)
+    - **Verification Before Any MFE/MAE Analysis:**
+      ```sql
+      -- Check if data is percentages or dollars
+      SELECT 
+        "entryPrice",
+        "maxFavorablePrice",
+        "maxFavorableExcursion" as stored,
+        CASE 
+          WHEN direction = 'long' THEN 
+            (("maxFavorablePrice" - "entryPrice") / "entryPrice") * 100
+          ELSE 
+            (("entryPrice" - "maxFavorablePrice") / "entryPrice") * 100
+        END as calculated_pct
+      FROM "Trade" 
+      WHERE "exitReason" IS NOT NULL 
+      ORDER BY "createdAt" DESC 
+      LIMIT 5;
+      
+      -- If stored ≠ calculated_pct → OLD DATA, use date filter
+      ```
+    - **See Also:**
+      * Common Pitfall #54 - MFE/MAE stored as dollars (supposedly fixed Nov 23)
+      * Revert commits: a15f17f "revert: Undo exit strategy optimization based on corrupted MFE data"
+      * Original bug commits: a67a338 (code), f65aae5 (docs)
+    - **Git commits:** a15f17f (revert), a67a338 (incorrect optimization), f65aae5 (incorrect docs)
+    - **Status:** ✅ Fixed - Analysis methodology documented, incorrect changes reverted
+
+73. **CRITICAL SECURITY: .env file tracked in git (CRITICAL - Fixed Dec 5, 2025 - PR #3):**
     - **Symptom:** Sensitive credentials exposed in git repository history
     - **Credentials exposed:**
       * Database connection strings (PostgreSQL)