Phase 1 complete: S3-S6 strategies, S4 variant analysis, learnings doc

- S3 Key Level Breakout: best performer (52-53% WR, PF ~1.0 on JPY crosses) - S4 EMA Ribbon: tested 7 variants (D/E/F/F-v2/G/G-Minimal), exhausted - Only EUR_AUD S4-F marginally profitable (PF 1.06) - Detailed filter funnel analysis revealed contradictory filter stacking - S5 Momentum Exhaustion: extended to 5 pairs, PF 0.43-0.77 - S6 EMA Bounce: 59-60% WR but PF 0.83-0.84, needs SL/TP restructuring - Added STRATEGY_LEARNINGS.md with design principles and next steps - Added M5 data downloader for 3-timeframe strategies - Updated README with full strategy scorecard Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-08-23 15:08:05 +00:00 · 2026-02-18 20:42:16 +10:00
parent dce54845c2
commit edbe359d1b
88 changed files with 12570 additions and 2963 deletions
@@ -0,0 +1,145 @@
+"""
+Strategy S4-F: EMA Ribbon — Trend Context Filter.
+
+Entry: Current S4 ribbon logic PLUS:
+  - 1H trend for LONG: 1H close > 1H 200 EMA AND 1H 50 EMA > 1H 200 EMA
+  - 1H trend for SHORT: 1H close < 1H 200 EMA AND 1H 50 EMA < 1H 200 EMA
+  - Price within 1.5 ATR of 1H 50 EMA
+  - 15min EMA stacking for LONG: 8 EMA > 13 EMA AND 13 EMA > 21 EMA
+  - 15min EMA stacking for SHORT: 8 EMA < 13 EMA AND 13 EMA < 21 EMA
+  - Volume > 1.2x average
+
+Exit: Same as S4-D (SL 2.0 ATR, TP 3.0 ATR, 100% close, no partials)
+"""
+from typing import Optional
+import numpy as np
+import pandas as pd
+from .base import BaseStrategy
+
+
+class S4F_EMA_Ribbon(BaseStrategy):
+    strategy_id = 4
+    name = "S4F_Trend_Context"
+
+    def check_signal(self, data: pd.DataFrame, idx: int,
+                     current: pd.Series,
+                     htf_row: Optional[pd.Series] = None) -> Optional[dict]:
+        if idx < 50:
+            return None
+
+        hour = current.name.hour if hasattr(current.name, 'hour') else 0
+        if hour < 8 or hour >= 16:
+            return None
+
+        atr_val = current.get("atr_14", 0)
+        if atr_val <= 0 or np.isnan(atr_val):
+            return None
+
+        if htf_row is None:
+            return None
+
+        # ------- ORIGINAL H1 EMA STACK (20 > 50 > 100 > 200) -------
+        htf_ema20 = htf_row.get("ema_20", np.nan)
+        htf_ema50 = htf_row.get("ema_50", np.nan)
+        htf_ema100 = htf_row.get("ema_100", np.nan)
+        htf_ema200 = htf_row.get("ema_200", np.nan)
+        if any(np.isnan(v) for v in [htf_ema20, htf_ema50, htf_ema100, htf_ema200]):
+            return None
+
+        long_stack = htf_ema20 > htf_ema50 > htf_ema100 > htf_ema200
+        short_stack = htf_ema20 < htf_ema50 < htf_ema100 < htf_ema200
+        if not long_stack and not short_stack:
+            return None
+
+        # ------- NEW 1H TREND FILTER -------
+        htf_close = htf_row.get("close", np.nan)
+        if np.isnan(htf_close):
+            return None
+
+        if long_stack:
+            if not (htf_close > htf_ema200 and htf_ema50 > htf_ema200):
+                return None
+            direction = "LONG"
+        else:
+            if not (htf_close < htf_ema200 and htf_ema50 < htf_ema200):
+                return None
+            direction = "SHORT"
+
+        # ------- PRICE WITHIN 1.5 ATR OF 1H 50 EMA -------
+        price = current["close"]
+        if abs(price - htf_ema50) > 1.5 * atr_val:
+            return None
+
+        # ------- M15 RIBBON COMPRESSION -> EXPANSION -------
+        ema_20 = current.get("ema_20", np.nan)
+        ema_50 = current.get("ema_50", np.nan)
+        ema_100 = current.get("ema_100", np.nan)
+        if any(np.isnan(v) for v in [ema_20, ema_50, ema_100]):
+            return None
+
+        ribbon_width = max(ema_20, ema_50, ema_100) - min(ema_20, ema_50, ema_100)
+        compression_threshold = 1.0 * atr_val
+
+        was_compressed = False
+        min_compression_width = float('inf')
+        for j in range(max(0, idx - 20), idx):
+            bar = data.iloc[j]
+            e20 = bar.get("ema_20", np.nan)
+            e50 = bar.get("ema_50", np.nan)
+            e100 = bar.get("ema_100", np.nan)
+            if any(np.isnan(v) for v in [e20, e50, e100]):
+                continue
+            w = max(e20, e50, e100) - min(e20, e50, e100)
+            if w <= compression_threshold:
+                was_compressed = True
+                min_compression_width = min(min_compression_width, w)
+
+        if not was_compressed:
+            return None
+
+        if ribbon_width <= min_compression_width * 1.2:
+            return None
+
+        # ------- 15MIN EMA STACKING: 8 > 13 > 21 (LONG) or reversed -------
+        ema_8 = current.get("ema_8", np.nan)
+        ema_13 = current.get("ema_13", np.nan)
+        ema_21 = current.get("ema_21", np.nan)
+        if any(np.isnan(v) for v in [ema_8, ema_13, ema_21]):
+            return None
+
+        if direction == "LONG":
+            if not (ema_8 > ema_13 and ema_13 > ema_21):
+                return None
+        else:
+            if not (ema_8 < ema_13 and ema_13 < ema_21):
+                return None
+
+        # ------- VOLUME > 1.2x average -------
+        if "volume" not in current.index:
+            return None
+        vol = current["volume"]
+        vol_avg = data["volume"].iloc[max(0, idx - 20):idx].mean()
+        if vol_avg <= 0 or vol <= 1.2 * vol_avg:
+            return None
+
+        # ------- EXIT LEVELS -------
+        if direction == "LONG":
+            sl = price - 2.0 * atr_val
+            tp1 = price + 3.0 * atr_val
+        else:
+            sl = price + 2.0 * atr_val
+            tp1 = price - 3.0 * atr_val
+
+        return {
+            "direction": direction,
+            "sl": sl,
+            "tp1": tp1,
+            "tp2": tp1,
+            "tp3": tp1,
+            "confluence": 3,
+            "entry_pattern": "ribbon_trend_context",
+            "tp_splits": (1.0, 0.0, 0.0),
+            "trail_atr_mult": 0,
+            "max_bars": 60,
+            "no_breakeven": True,
+        }