feat: grade gate-ablation variants under the hold-to-horizon exit too

The ablation judged floors under the target/stop model, but the exit
sweeps point at replacing that exit with a fixed hold — under which the
R:R floor's rationale (bigger payoff at the target) may not apply. Each
ablation row now also carries hold_avg_r / hold_net_avg_r / hold_total_r
(30d hold, initial stop only), so the Phase 3 gate decision can be read
under the exit policy that would actually be used.

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

app/services/backtest_service.py

@@ -911,16 +911,27 @@ def _gate_ablation(candidates: list[dict], activation: dict, threshold: float) -
         ("no_neutral_exclusion", [rr_ok, conf_ok, tighteners_ok]),
         ("momentum_only", []),
     ]
-    return [
-        {
+    # Grade each variant under BOTH exit models: the target/stop outcome
+    # (_bucket_stats) and the hold-to-horizon time exit. A floor that pays under
+    # the target model may be meaningless once the exit is a fixed hold — the
+    # hold_* columns are what a time-exit gate decision should read.
+    hold_days = max(TIME_EXIT_DAYS)
+    rows: list[dict] = []
+    for name, checks in variants:
+        matching = [
+            c for c in candidates
+            if momentum_ok(c) and all(check(c) for check in checks)
+        ]
+        hold = _time_exit_bucket(matching, hold_days)
+        rows.append({
             "variant": name,
-            **_bucket_stats([
-                c for c in candidates
-                if momentum_ok(c) and all(check(c) for check in checks)
-            ]),
-        }
-        for name, checks in variants
-    ]
+            **_bucket_stats(matching),
+            "hold_days": hold_days,
+            "hold_avg_r": hold["avg_r"],
+            "hold_net_avg_r": hold["net_avg_r"],
+            "hold_total_r": hold["total_r"],
+        })
+    return rows
 
 
 async def run_backtest(
@@ -1051,7 +1062,10 @@ async def run_backtest(
             "Each row re-qualifies the same candidates at the current momentum "
             f"cutoff ({current_min_pct:.0f}) with one floor removed (long-only "
             "while the momentum gate is active). If dropping a floor doesn't "
-            "hurt net expectancy, that floor isn't pulling its weight."
+            "hurt net expectancy, that floor isn't pulling its weight. The Hold "
+            "columns grade the same variants under the hold-to-horizon time exit "
+            "instead of the S/R target — the view that matters if the exit "
+            "policy moves to a fixed hold."
         ),
         "take_profit_sweep": [_take_profit_bucket(qualified, tp) for tp in TP_LEVELS],
         "trailing_sweep": [_trailing_bucket(qualified, round(f * 100)) for f in TRAIL_LEVELS],