fix: align production defaults and close review parity gaps

Ship greenfield min_rr=2.0 and conf=0, read-only Structural S/R, indicator
cache invalidation, and UI/gate language that treats GTL as screening not exit.
Align strategy_rank missing-vol fallback live vs backtest, single-source
PRIMARY_TARGET_MIN_RR, expand prod parity tests, and drop dead FE clients.

app/services/backtest_service.py

@@ -59,6 +59,7 @@ from app.services.admin_service import get_activation_config, update_setting
 from app.services.indicator_service import _extract_ohlcv, compute_atr
 from app.services.momentum_service import (
     STRATEGY_RANK_MOMENTUM_WEIGHT,
+    blend_strategy_rank,
     compute_realized_vol_6m,
 )
 from app.services.outcome_service import (
@@ -77,6 +78,7 @@ from app.services.qualification import (
     setup_qualifies,
 )
 from app.services.recommendation_service import (
+    PRIMARY_TARGET_MIN_RR,
     _choose_recommended_action,
     _classify_by_probability,
     _prune_floor_pinned_targets,
@@ -337,11 +339,11 @@ def _window_setups(
         targets = _prune_floor_pinned_targets(targets)
         primary = _select_primary_target(
             targets,
-            min_rr=1.5,
+            min_rr=PRIMARY_TARGET_MIN_RR,
         )
         if primary is None:
             continue
-        # Flag the primary so qualification's EV uses the primary target's
+        # Flag the primary so qualification uses the primary target's
         # probability (matching production's enhance_trade_setup).
         for t in targets:
             t["is_primary"] = t is primary
@@ -1265,15 +1267,27 @@ def _assign_weighted_blend(
     primary_weight: float,
     secondary_key: str,
 ) -> None:
-    secondary_weight = 1.0 - primary_weight
+    """Blend ranks; fall back to primary when secondary is missing.
+
+    Matches live ``blend_strategy_rank`` for the production 80/20 key: a name
+    with residual momentum but no vol history keeps its mom percentile instead
+    of ranking as 0 / None at the bottom of the book.
+    """
     for c in candidates:
         primary = c.get(primary_key)
         secondary = c.get(secondary_key)
-        c[output_key] = (
-            primary * primary_weight + secondary * secondary_weight
-            if primary is not None and secondary is not None
-            else None
-        )
+        # Production weight path: reuse the shared helper so live/sim cannot drift.
+        if primary_weight == STRATEGY_RANK_MOMENTUM_WEIGHT:
+            c[output_key] = blend_strategy_rank(
+                None if primary is None else float(primary),
+                None if secondary is None else float(secondary),
+                momentum_weight=primary_weight,
+            )
+            continue
+        if primary is not None and secondary is not None:
+            c[output_key] = primary * primary_weight + secondary * (1.0 - primary_weight)
+        else:
+            c[output_key] = primary
 
 
 def _assign_residual_low_vol_blend(candidates: list[dict]) -> None: