Dev

README.md

@@ -38,6 +38,11 @@ historical reproduction.
 - Explicit margin, leverage, fees, slippage, funding, and liquidation handling.
 - Stable audit artifacts: metrics, plots, raw reports, trade logs, config JSON,
   run manifest, and optional QuantStats HTML.
+- Nautilus certification bundles with parity CSVs, tolerance profiles, known
+  differences, and explicit skip/pass status for optional trustee workflows.
+- Nautilus package-depth preflight with OHLCV volume caps and synthetic-book
+  stress tests for spread, queue, participation, partial-fill, and package
+  rejection assumptions.
 - Walk-forward and train/test optimization designed to avoid leaking OOS data
   into parameter selection, plus full-sample robust calibration for final
   production parameter discovery.
@@ -88,6 +93,21 @@ Cython/C++ work. See `benchmarks/phase9_optimization_report.md`,
 `benchmarks/portfolio_real_parity_report.md` for parity and optimization
 history.
 
+Latest Phase 14C service-loop optimization benchmark:
+
+| Workload | Cold / full route | Prepared / minimal route | Speedup | Parity |
+|---|---:|---:|---:|---|
+| Single-symbol WFO | 1.007s | 0.883s | 1.14x | pass |
+| Portfolio WFO | 0.440s | 0.326s | 1.35x | pass |
+| Native-event order replay | 0.0096s | 0.0046s | 2.10x | pass |
+| Arbitrage package replay | 0.0746s | 0.0663s | 1.13x | pass |
+| Native portfolio reports | 0.0410s full | 0.0233s minimal | 1.76x | pass |
+
+Phase 14C added run-local prepared market-array reuse for WFO/service loops and
+`report_level="full" | "standard" | "minimal"` for native portfolio reports.
+The default remains `full`; lighter report levels are opt-in for optimizers and
+services, and parity tests lock core accounting equality before any speed claim.
+
 Ecosystem positioning:
 
 | Tool | Core strength | Runtime model | QuantBT role beside it |
@@ -212,6 +232,12 @@ Experimental Nautilus package validation is available for DCA/grid,
 bracket/OCO, basket, and portfolio workflows by compiling strategy state into
 explicit order packages.
 
+Package-depth validation is opt-in. `depth_model="ohlcv_volume_cap"` is the
+default Level-1 preflight, `depth_model="synthetic_book"` creates deterministic
+Level-2 stress books from spread/depth assumptions, and future
+`depth_model="l2_replay"` is intentionally gated until real venue snapshots,
+incremental updates, and trade prints are provided.
+
 Supported Binance perpetual validation instruments:
 
 `BTCUSDT-PERP.BINANCE`, `ETHUSDT-PERP.BINANCE`, `BNBUSDT-PERP.BINANCE`,

__init__.py

@@ -92,6 +92,8 @@
 from .core.execution_depth import (
     NautilusExecutionDepthConfig,
     PackageDepthPreflightResult,
+    SUPPORTED_DEPTH_MODELS,
+    l2_replay_available,
     simulate_nautilus_order_package_depth,
 )
 from .core.structured_orders import (
@@ -194,6 +196,7 @@
 from .reporting import (
     build_arbitrage_domain_audit,
     build_native_nautilus_parity_report,
+    build_nautilus_certification_profile,
     build_nautilus_depth_execution_report,
     build_nautilus_depth_parity_summary,
     build_nautilus_pct_equity_diagnostic,
@@ -202,7 +205,9 @@
     build_portfolio_nautilus_validation_report,
     compare_native_arbitrage_results,
     export_nautilus_report_bundle,
+    NautilusToleranceProfile,
     summarize_native_nautilus_parity_report,
+    write_nautilus_certification_artifacts,
 )
 
 __version__ = "0.1.0"
@@ -237,6 +242,7 @@
     "NATIVE_PORTFOLIO_SUPPORTED_SIZING_MODES",
     "build_arbitrage_domain_audit",
     "build_native_nautilus_parity_report",
+    "build_nautilus_certification_profile",
     "build_nautilus_depth_execution_report",
     "build_nautilus_depth_parity_summary",
     "build_nautilus_pct_equity_diagnostic",
@@ -245,7 +251,9 @@
     "build_portfolio_nautilus_validation_report",
     "compare_native_arbitrage_results",
     "export_nautilus_report_bundle",
+    "NautilusToleranceProfile",
     "summarize_native_nautilus_parity_report",
+    "write_nautilus_certification_artifacts",
     "WalkForwardConfig",
     "WalkForwardEngine",
     "WalkForwardFold",
@@ -343,6 +351,8 @@
     "portfolio_capability_matrix",
     "quantize_signed_quantity",
     "round_down_to_step",
+    "SUPPORTED_DEPTH_MODELS",
+    "l2_replay_available",
     "simulate_nautilus_order_package_depth",
     "validate_portfolio_result_contract",
     # metrics

benchmarks/phase15a_nautilus_certification.json

@@ -0,0 +1,41 @@
+{
+  "diff_workflows": 0,
+  "failed_workflows": 0,
+  "include_nautilus": false,
+  "output_dir": "/root/bobby/pool_alpha/quantbt/benchmarks/phase15a_nautilus_bundles",
+  "passed_workflows": 0,
+  "skipped_workflows": 5,
+  "status": "pass",
+  "workflows": [
+    {
+      "bundle_dir": null,
+      "reason": "run with --include-nautilus",
+      "status": "skipped",
+      "workflow": "pct_equity_signal"
+    },
+    {
+      "bundle_dir": null,
+      "reason": "run with --include-nautilus",
+      "status": "skipped",
+      "workflow": "explicit_orders"
+    },
+    {
+      "bundle_dir": null,
+      "reason": "run with --include-nautilus",
+      "status": "skipped",
+      "workflow": "basket_package"
+    },
+    {
+      "bundle_dir": null,
+      "reason": "run with --include-nautilus",
+      "status": "skipped",
+      "workflow": "portfolio_package"
+    },
+    {
+      "bundle_dir": null,
+      "reason": "run with --include-nautilus",
+      "status": "skipped",
+      "workflow": "basis_arbitrage_package"
+    }
+  ]
+}

benchmarks/phase15a_nautilus_certification.md

@@ -0,0 +1,35 @@
+# Phase 15A Nautilus Certification Bundles
+
+Status: **pass**
+
+- Include Nautilus: `False`
+- Output directory: `/root/bobby/pool_alpha/quantbt/benchmarks/phase15a_nautilus_bundles`
+- Passed workflows: `0`
+- Skipped workflows: `5`
+- Failed workflows: `0`
+
+## Workflow Matrix
+
+| workflow | status | bundle | tolerance status | reason |
+| --- | --- | --- | --- | --- |
+| `pct_equity_signal` | `skipped` | `` | `` | run with --include-nautilus |
+| `explicit_orders` | `skipped` | `` | `` | run with --include-nautilus |
+| `basket_package` | `skipped` | `` | `` | run with --include-nautilus |
+| `portfolio_package` | `skipped` | `` | `` | run with --include-nautilus |
+| `basis_arbitrage_package` | `skipped` | `` | `` | run with --include-nautilus |
+
+## Required Bundle Files
+
+- `config.json`
+- `run_manifest.json`
+- `metrics_summary.json`
+- `equity_curve.csv`, `returns.csv`, `account_report.csv`
+- `orders_report.csv`, `fills_report.csv`, `positions_report.csv`
+- `trade_log.csv`, `fill_log.txt`
+- `native_vs_nautilus_parity.csv`
+- `tolerance_profile.json`
+- `known_differences.md`
+
+## Interpretation
+
+A skipped workflow is not a pass claim. It means the optional Nautilus dependency or instrument route was not available in this environment. A pass means the workflow produced a bundle and satisfied the declared tolerance profile.

benchmarks/phase15b_synthetic_depth.json

@@ -0,0 +1,38 @@
+{
+  "phase": "15B",
+  "status": "pass",
+  "l2_replay_available": false,
+  "cases": [
+    {
+      "case": "synthetic_market_vwap",
+      "depth_model": "synthetic_book",
+      "status": "filled",
+      "filled_qty": 2.0,
+      "fill_price": 100.1,
+      "levels_consumed": 2,
+      "accepted_orders": 1,
+      "rejected_orders": 0
+    },
+    {
+      "case": "synthetic_partial_queue",
+      "depth_model": "synthetic_book",
+      "status": "partial",
+      "filled_qty": 1.5,
+      "fill_price": 100.02333333333333,
+      "levels_consumed": 2,
+      "accepted_orders": 1,
+      "rejected_orders": 0
+    },
+    {
+      "case": "ohlcv_all_or_none_baseline",
+      "depth_model": "ohlcv_volume_cap",
+      "status": "filled",
+      "filled_qty": 1.0,
+      "fill_price": 100.0,
+      "levels_consumed": 1,
+      "accepted_orders": 1,
+      "rejected_orders": 0
+    }
+  ],
+  "claim_scope": "Level-2 synthetic stress only; not venue L2 replay."
+}

benchmarks/phase15b_synthetic_depth.md

@@ -0,0 +1,16 @@
+# Phase 15B Synthetic Depth Evidence
+
+Status: **pass**
+
+- L2 replay provider available: `False`
+- Claim scope: Level-2 synthetic stress only; not venue L2 replay.
+
+| case | depth model | status | filled qty | fill price | levels | accepted | rejected |
+| --- | --- | --- | ---: | ---: | ---: | ---: | ---: |
+| `synthetic_market_vwap` | `synthetic_book` | `filled` | 2.00000000 | 100.10000000 | 2 | 1 | 0 |
+| `synthetic_partial_queue` | `synthetic_book` | `partial` | 1.50000000 | 100.02333333 | 2 | 1 | 0 |
+| `ohlcv_all_or_none_baseline` | `ohlcv_volume_cap` | `filled` | 1.00000000 | 100.00000000 | 1 | 1 | 0 |
+
+## Interpretation
+
+Synthetic depth proves deterministic queue, participation, spread and level-consumption behavior. It does not certify real exchange queue priority. Real L2 certification remains gated by venue snapshots, incremental updates and trade prints.