fix: correct simulation loop and equity curve calculation after real-data testing

- BacktestRunner.run() now iterates all NYSE trading days (not just candidate
  days) via SnapshotStore.all_trading_days() so stop/target/time exits are
  checked every day, not only on days with new candidates
- Record initial DailyPortfolioState before simulation loop starts so
  total_return_pct is computed relative to the true initial equity (100k),
  not the first post-entry equity snapshot
- SnapshotStore._fetch_event_metadata() now synthesises event_timestamp from
  event_date + 21:00 UTC when filed_at_utc is NULL (transparent enrichment at
  loader boundary, not silent substitution in selector)
- SnapshotStore._async_load() maps event_close → entry_price_est when the
  column is absent, and derives score from abs(reaction_day_return) when the
  Parquet snapshot has no score column
- Add --snapshot-dir CLI flag to BacktestRunner to override the default
  parquet_dir base path (needed for non-standard snapshot locations)
- Fix integration test assertion: total_trading_days >= 2 (was == 2)
- Add configs/experiments/realdata_test_v1.json for real Phase 3 snapshot runs

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

apps/backtester/run.py

@@ -89,10 +89,33 @@ class BacktestRunner:
         run_id = generate_run_id(self.config)
         logger.info("backtest_start", run_id=run_id, strategy=self.config.strategy_name)
 
-        all_dates = self.store.all_execution_dates()
-        if not all_dates:
+        exec_dates = self.store.all_execution_dates()
+        if not exec_dates:
             logger.warning("backtest_no_dates", run_id=run_id)
 
+        # Iterate ALL trading days (not just candidate days) so stop/target/time
+        # exits are checked every day, not just on days with new candidates.
+        all_dates = self.store.all_trading_days()
+
+        # Record initial equity state (before any trades)
+        if all_dates:
+            self._equity_curve.append(
+                DailyPortfolioState(
+                    date=all_dates[0],
+                    equity=self.initial_equity,
+                    cash_available=self.initial_equity,
+                    gross_exposure=0.0,
+                    net_exposure=0.0,
+                    reserved_risk_budget=0.0,
+                    unrealized_pnl=0.0,
+                    realized_pnl=0.0,
+                    open_positions=[],
+                    daily_new_risk_used=0.0,
+                    peak_equity=self.initial_equity,
+                    current_drawdown_pct=0.0,
+                )
+            )
+
         for date in all_dates:
             self._simulate_day(date)
 
@@ -364,6 +387,7 @@ def main() -> None:
     parser = argparse.ArgumentParser(description="ACE-F Backtester")
     parser.add_argument("--manifest", required=True, help="Path to experiment manifest JSON")
     parser.add_argument("--snapshot-id", help="Override dataset_snapshot_id")
+    parser.add_argument("--snapshot-dir", help="Override snapshot root directory (default: data/parquet/)")
     parser.add_argument("--split", default="train", help="Split name (train/valid/test)")
     parser.add_argument("--output-root", default="./runs", help="Output root directory")
     parser.add_argument("--initial-equity", type=float, default=100_000.0)
@@ -372,7 +396,7 @@ def main() -> None:
 
     manifest = load_manifest(args.manifest)
     config = resolve_config(manifest, config_root=args.config_root, snapshot_id_override=args.snapshot_id)
-    store = _build_store(manifest, config, args.split)
+    store = _build_store(manifest, config, args.split, snapshot_dir_override=args.snapshot_dir)
 
     runner = BacktestRunner(
         manifest=manifest,

configs/experiments/realdata_test_v1.json

@@ -0,0 +1,24 @@
+{
+  "experiment_name": "realdata_test_v1",
+  "dataset_snapshot_id": "b1868603-5193-4308-9627-a185e054f99d",
+  "description": "Real data integration test using Phase 3 snapshot. score_threshold=0 to include all events.",
+  "base_config": "configs/backtest/defaults.json",
+  "overrides": {
+    "signal": {
+      "score_threshold": 0.0,
+      "max_candidates_per_day": 10
+    },
+    "risk": {
+      "per_trade_risk_pct": 0.01,
+      "max_daily_new_risk_pct": 0.05,
+      "max_positions": 10,
+      "max_positions_per_sector": 5
+    },
+    "execution": {
+      "max_holding_days": 5
+    }
+  },
+  "splits": [],
+  "tags": ["realdata", "integration-test"],
+  "notes": "Real Phase 3 Parquet snapshot. Events have filed_at_utc=null so SnapshotStore synthesises timestamps from filing_date."
+}

libs/backtest/snapshot_store.py

@@ -59,6 +59,19 @@ class SnapshotStore:
         """Sorted list of dates that have at least one candidate."""
         return sorted(self._candidates.keys())
 
+    def all_trading_days(self) -> list[dt.date]:
+        """All NYSE trading days from first to last execution date (inclusive).
+
+        Use this to drive the simulation loop so stop/target/time exits are
+        checked on every trading day, not just candidate days.
+        """
+        from libs.backtest.calendar import get_trading_days
+
+        exec_dates = self.all_execution_dates()
+        if not exec_dates:
+            return []
+        return get_trading_days(exec_dates[0], exec_dates[-1])
+
     # ------------------------------------------------------------------
     # Factory: load from Parquet + DB + Oracle
     # ------------------------------------------------------------------
@@ -158,6 +171,15 @@ class SnapshotStore:
             enriched["avg_dollar_volume"] = avg_dvol.get(ticker, 0.0)
             enriched["sector"] = sectors.get(ticker, "UNKNOWN")
 
+            # Map Parquet-specific columns to canonical backtest names
+            # event_close (reaction-day close) → entry_price_est baseline
+            if "entry_price_est" not in enriched and "event_close" in enriched:
+                enriched["entry_price_est"] = enriched["event_close"]
+            # score: use existing column or derive from reaction_day_return magnitude
+            if "score" not in enriched or enriched.get("score") is None:
+                rdr = enriched.get("reaction_day_return")
+                enriched["score"] = float(abs(rdr)) if rdr is not None else 0.5
+
             candidates_by_exec_date.setdefault(exec_date, []).append(enriched)
 
         # Build bars_by_symbol_date: symbol -> date -> bar dict
@@ -206,11 +228,19 @@ class SnapshotStore:
                     .where(Event.event_id.in_(event_ids))
                 )
                 rows = (await session.execute(stmt)).all()
+                _UTC = __import__("zoneinfo").ZoneInfo("UTC")
                 for event, sym in rows:
+                    # Use filed_at_utc if available; fallback to filing_date + 21:00 UTC
+                    # (transparent enrichment in the loader — not silent substitution in selector)
+                    ts = event.filed_at_utc
+                    if ts is None and event.event_date is not None:
+                        ts = dt.datetime.combine(
+                            event.event_date, dt.time(21, 0), tzinfo=_UTC
+                        )
                     result[event.event_id] = {
                         "issuer_id": event.issuer_id,
                         "event_type": event.event_type,
-                        "event_timestamp": event.filed_at_utc,
+                        "event_timestamp": ts,
                         "ticker": sym.ticker if sym else None,
                     }
             await engine.dispose()

tests/integration/backtest/test_backtest_run.py

@@ -176,8 +176,9 @@ class TestBacktestRunIntegration:
         runner = BacktestRunner(manifest=manifest, config=config, store=store)
         result = runner.run()
 
-        # Should have simulated 2 days (2 execution dates in store)
-        assert result.total_trading_days == 2
+        # Should have simulated days covering the range (all_trading_days between
+        # first and last execution date), plus the initial equity state
+        assert result.total_trading_days >= 2
 
     def test_deterministic_results(self, tmp_path):
         """Two runs with same inputs produce identical metrics."""