Activation gate: primary target probability floor (>= 20%)

A qualified setup's primary target must now clear MIN_TARGET_PROBABILITY
(20%), shared with the primary-selection floor in recommendation_service
and mirrored in the frontend gate. Closes the read-time hole where a
stale pre-c7a198b row starring a far lottery target (probability pinned
at the 3% clamp floor, R:R inflated by the same distance) qualified
forever: the scanner emits no replacement row and live R:R never decays.

A/B backtest vs c7a198b baseline (same July-3 snapshot): 7 of 1096
qualified setups removed; qualified net avg R 0.202 -> 0.207, hold
Sharpe 2.00 -> 2.02, CAGR +48.8% -> +49.6%, max DD unchanged at -15.8%.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

app/services/qualification.py

@@ -5,9 +5,11 @@ performance stats (server) and mirrored on the frontend. The core selection is
 residual cross-sectional momentum: a setup's ticker must rank in the top
 ``min_momentum_percentile`` of the universe by beta-adjusted 12-1 month momentum.
 R:R and confidence remain as floors, and conviction/conflict survive as optional
-tighteners (off by default). Qualified setups must also have a probability-backed
-target; otherwise a mathematically high R:R can be driven by a fragile target
-with no independent validation.
+tighteners (off by default). Qualified setups must also have a primary target
+with at least ``MIN_TARGET_PROBABILITY`` reach probability: a primary below the
+floor is a lottery target whose distance inflates R:R, so it would otherwise
+game the min_rr gate (the model clamps probabilities at 3%, and far targets pin
+there while their live R:R stays high forever).
 """
 
 from __future__ import annotations
@@ -16,6 +18,13 @@ from typing import Any
 
 HIGH_CONVICTION_ACTIONS = {"LONG_HIGH", "SHORT_HIGH"}
 
+# Floor for the primary target's reach probability, shared with the primary
+# target selection in recommendation_service and mirrored in the frontend
+# (qualification.ts). Under the two-barrier model a fair-race 1.5:1 target sits
+# near ~34% before drift adjustments, so 20% only excludes targets the model
+# itself considers long shots.
+MIN_TARGET_PROBABILITY = 20.0
+
 
 def _action_direction(action: str | None) -> str:
     if not action or action == "NEUTRAL":
@@ -85,7 +94,8 @@ def setup_qualifies(setup: Any, config: dict) -> bool:
         live_rr = live_risk_reward(setup, float(current_price))
         if live_rr is not None and live_rr < config["min_rr"]:
             return False
-    if primary_target_probability(setup) is None:
+    target_probability = primary_target_probability(setup)
+    if target_probability is None or target_probability < MIN_TARGET_PROBABILITY:
         return False
     if (setup.confidence_score or 0.0) < config["min_confidence"]:
         return False

app/services/recommendation_service.py

@@ -13,6 +13,7 @@ from app.models.settings import SystemSetting
 from app.models.sr_level import SRLevel
 from app.models.ticker import Ticker
 from app.models.trade_setup import TradeSetup
+from app.services.qualification import MIN_TARGET_PROBABILITY
 from app.services.sr_service import cluster_sr_zones
 
 logger = logging.getLogger(__name__)
@@ -575,10 +576,10 @@ def build_recommendation_snapshot(
 
 
 PRIMARY_TARGET_MIN_RR = 1.5
-# Below this the target is a lottery ticket: under the two-barrier model a
-# fair-race 1.5:1 target sits near ~34% before drift adjustments, so 20% only
-# excludes targets the model itself considers long shots.
-PRIMARY_TARGET_MIN_PROBABILITY = 20.0
+# Below this the target is a lottery ticket. Shared with the activation gate
+# (qualification.MIN_TARGET_PROBABILITY) so the primary selection and the gate
+# agree on what counts as a probability-backed target.
+PRIMARY_TARGET_MIN_PROBABILITY = MIN_TARGET_PROBABILITY
 
 
 def _select_primary_target(

frontend/src/lib/qualification.ts

@@ -2,6 +2,13 @@ import type { ActivationConfig, TradeSetup } from './types';
 
 const HIGH_CONVICTION_ACTIONS = new Set(['LONG_HIGH', 'SHORT_HIGH']);
 
+/**
+ * Floor for the primary target's reach probability — mirrors
+ * MIN_TARGET_PROBABILITY in app/services/qualification.py. A primary below
+ * this is a lottery target whose distance inflates R:R past the min_rr gate.
+ */
+export const MIN_TARGET_PROBABILITY = 20;
+
 function actionDirection(action: TradeSetup['recommended_action']): 'long' | 'short' | 'neutral' {
   if (!action || action === 'NEUTRAL') return 'neutral';
   if (action.startsWith('LONG')) return 'long';
@@ -40,7 +47,7 @@ export function qualifiesSetup(setup: TradeSetup, config: ActivationConfig): boo
     return false;
   }
   const targetProbability = primaryTargetProbability(setup);
-  if (targetProbability == null || targetProbability <= 0) return false;
+  if (targetProbability == null || targetProbability < MIN_TARGET_PROBABILITY) return false;
   if ((setup.confidence_score ?? 0) < config.min_confidence) return false;
   // Residual cross-sectional momentum is the core selection (long-only). While
   // the gate is active, shorts never qualify; missing ranks do not qualify
@@ -77,6 +84,9 @@ export function disqualifyReason(setup: TradeSetup, config: ActivationConfig): s
   }
   const targetProbability = primaryTargetProbability(setup);
   if (targetProbability == null || targetProbability <= 0) return 'no target probability';
+  if (targetProbability < MIN_TARGET_PROBABILITY) {
+    return `target probability below ${MIN_TARGET_PROBABILITY}%`;
+  }
   if ((setup.confidence_score ?? 0) < config.min_confidence) {
     return `confidence below ${config.min_confidence.toFixed(0)}%`;
   }

tests/unit/test_qualification.py

@@ -82,6 +82,31 @@ class TestFloors:
             DEFAULT_GATE,
         ) is True
 
+    def test_lottery_primary_target_fails(self):
+        # A far target pinned at the model's 3% clamp floor: its distance keeps
+        # both stored and live R:R above the gate, so only the probability floor
+        # can reject it (the stale pre-fix lottery-headline case).
+        s = _setup(rr_ratio=3.09, targets=[{"probability": 3.0, "is_primary": True}])
+        assert setup_qualifies(s, DEFAULT_GATE) is False
+
+    def test_probability_at_floor_passes(self):
+        assert setup_qualifies(
+            _setup(targets=[{"probability": 20.0, "is_primary": True}]),
+            DEFAULT_GATE,
+        ) is True
+
+    def test_probability_just_below_floor_fails(self):
+        assert setup_qualifies(
+            _setup(targets=[{"probability": 19.9, "is_primary": True}]),
+            DEFAULT_GATE,
+        ) is False
+
+    def test_best_target_fallback_below_floor_fails(self):
+        # No starred primary: the fallback takes the best target, which must
+        # still clear the probability floor.
+        s = _setup(targets=[{"probability": 12.0}, {"probability": 8.0}])
+        assert setup_qualifies(s, DEFAULT_GATE) is False
+
 
 class TestMomentumGate:
     def test_top_momentum_passes(self):