Stop-Loss Validation

Core Rules

Note: Use PRICE_EPSILON = entry_price * 1e-8 for floating-point comparison.

Fixed Bug

Commit: 7f940fb Problem: When market moves fast, support/resistance levels may be on wrong side of entry price, causing immediate SL trigger.

Example:

Entry: $91,626 (LONG)
Support: $91,808 (above entry!)
Original: SL = $91,808 × 0.999 = $91,808.10
Result: SL triggered immediately, loss -$0.18 in 820ms

Fix:

# strategy/deepseek_strategy.py lines 1502-1543
PRICE_EPSILON = max(entry_price * 1e-8, 1e-8)  # Relative tolerance

if side == OrderSide.BUY:
    default_sl = entry_price * 0.98  # Default 2% SL
    if self.sl_use_support_resistance and support > 0:
        potential_sl = support * (1 - self.sl_buffer_pct)
        # Validate: SL must be below entry (with epsilon tolerance)
        if potential_sl < entry_price - PRICE_EPSILON:
            stop_loss_price = potential_sl
        else:
            stop_loss_price = default_sl  # Fallback to default 2%
            self.log.warning(f"⚠️ Support above entry, using default SL")

Test Command

cd /home/linuxuser/nautilus_AItrader
source venv/bin/activate
python3 test_sl_fix.py

Expected Output

============================================================
  Stop-Loss Fix Validation Test
============================================================
Test 1: Bug scenario: Support above entry ✅ Passed
Test 2: Normal scenario: Support below entry ✅ Passed
...
🎉 All tests passed! Stop-loss fix is correct!

Key Files

Verification Steps

1. Read strategy/deepseek_strategy.py lines 1502-1602
2. Confirm stop-loss validation logic exists (with PRICE_EPSILON tolerance)
3. Run python3 test_sl_fix.py
4. All tests pass = fix is correct