feat: add backtest module with DB cache and scenario comparison

Backtest improvements:
- Add backtest.py with Oracle DB-backed OHLCV cache (no repeated API calls)
- Add backtest_trades table to cache simulation results by params hash
  (same params -> instant load, skip re-simulation)
- Add walk-forward scenario comparison (--walkforward-cmp)
- Add trend ceiling filter (--trend-cmp, max gain threshold)
- Add ticker win-rate filter (--ticker-cmp, SQL-based instant analysis)
- Precompute daily_features once per data load (not per scenario)

Live bot fixes:
- monitor: add hard stop-loss from buy price (in addition to trailing)
- strategy: fix re-entry condition to require +1% above last sell price
- price_collector: add 48h backfill on startup for trend calculation
- main: call backfill_prices() at startup

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

core/monitor.py

@@ -19,7 +19,7 @@ TIME_STOP_MIN_GAIN_PCT = float(os.getenv("TIME_STOP_MIN_GAIN_PCT", "3"))
 
 
 def _check_trailing_stop(ticker: str, pos: dict, current: float) -> bool:
-    """트레일링 스탑 체크. 매도 시 True 반환."""
+    """트레일링 스탑(최고가 기준) + 고정 스탑(매수가 기준) 체크. 매도 시 True 반환."""
     trader.update_peak(ticker, current)
 
     pos = trader.get_positions().get(ticker)
@@ -27,15 +27,24 @@ def _check_trailing_stop(ticker: str, pos: dict, current: float) -> bool:
         return False
 
     peak = pos["peak_price"]
+    buy_price = pos["buy_price"]
     drop_from_peak = (peak - current) / peak
+    drop_from_buy = (buy_price - current) / buy_price  # 구매가 대비 하락률
 
     if drop_from_peak >= STOP_LOSS_PCT:
         reason = (
             f"트레일링스탑 | 최고가={peak:,.0f}원 → "
             f"현재={current:,.0f}원 ({drop_from_peak:.1%} 하락)"
         )
-        trader.sell(ticker, reason=reason)
-        return True
+        return trader.sell(ticker, reason=reason)
+
+    if drop_from_buy >= STOP_LOSS_PCT:
+        reason = (
+            f"스탑로스 | 매수가={buy_price:,.0f}원 → "
+            f"현재={current:,.0f}원 ({drop_from_buy:.1%} 하락)"
+        )
+        return trader.sell(ticker, reason=reason)
+
     return False
 
 
@@ -70,15 +79,17 @@ def _check_position(ticker: str, pos: dict) -> None:
     if current is None:
         return
 
-    pnl = (current - pos["buy_price"]) / pos["buy_price"] * 100
+    buy_price = pos["buy_price"]
+    pnl = (current - buy_price) / buy_price * 100
     peak = pos["peak_price"]
     drop_from_peak = (peak - current) / peak
+    drop_from_buy = (buy_price - current) / buy_price
     entry_time = pos.get("entry_time", datetime.now())
     elapsed_hours = (datetime.now() - entry_time).total_seconds() / 3600
 
     logger.info(
-        f"[감시] {ticker} 현재={current:,.0f} | 최고={peak:,.0f} | "
-        f"하락={drop_from_peak:.1%} | 수익률={pnl:+.1f}% | "
+        f"[감시] {ticker} 현재={current:,.0f} | 매수가={buy_price:,.0f} | 최고={peak:,.0f} | "
+        f"수익률={pnl:+.1f}% | peak하락={drop_from_peak:.1%} | buy하락={drop_from_buy:.1%} | "
         f"보유={elapsed_hours:.1f}h"
     )
 
@@ -94,7 +105,7 @@ def run_monitor(interval: int = CHECK_INTERVAL) -> None:
     """전체 포지션 감시 루프."""
     logger.info(
         f"모니터 시작 | 체크={interval}초 | "
-        f"트레일링스탑={STOP_LOSS_PCT:.0%} | "
+        f"트레일링스탑={STOP_LOSS_PCT:.1%} | "
         f"타임스탑={TIME_STOP_HOURS:.0f}h/{TIME_STOP_MIN_GAIN_PCT:+.0f}%"
     )
     while True: