Incentive monitor (market + non-market signals) and a preregistered paper forward study

[seathatflowsinourveins]

What

• blueprints/us-equities/incentive-monitor/monitor.py is a continuous, data-only recorder. It never places orders. It records:
  • SIP snapshots of every active tradable US equity (REST, about 24 calls per sweep against the measured 10,000/min data limit);
  • OPRA option trades for every contract (websocket *, msgpack), aggregated per root, with large prints kept;
  • Benzinga news (websocket *);
  • SEC EDGAR current filings (8-K, 6-K, Schedule 13D, SC TO-T, 425, S-1, 424B4; declared contact sent to SEC only);
  • Nasdaq Trader halts.
• The monitor also:
  • builds an unvalidated early-stage board with fixed weights;
  • publishes incentive_symbols and its own code hash, start time and stream downtime;
  • after a restart, rebuilds the session's state from its own files.
• forward-protocol-v1.json plus board_scan.py form a paper forward study, incentive-board-forward-v1-20260924, frozen before its first order. Its design:
  • decisions at 10:30 and 13:30 ET;
  • at most 5 operating companies with a non-price incentive and a fresh-snapshot gain of 0–10%;
  • 4 incentive-free controls per selected name from the same gain bucket, never shared across the day;
  • an exclusive decision ledger (no re-rolls), with a record for every refusal;
  • sha256 bindings for the code, the running monitor, the engine files, the pinned configs and the scan;
  • a single analysis point with the bootstrap and RNG fixed.
• Execution goes through the merged adaptive-paper mover trial mode (Mover trial mode for the adaptive NautilusTrader paper engine (SYN-verified; paper needs an isolated account) #172): exit X2, 200 USD per entry, 1x, and its own state root.

Review trail (all before any order)

1. Pre-outcome review of 920edde3: no-go, 19 findings. Resolved in 532e4b5f.
2. Re-review: no-go, N1–N9. Resolved in 1fba0f36 and 32a0c205.
3. Final check: go, with two conditions: the running monitor's code hash is bound, and the engine runs only on the recorded scan hash. Both are enforced in 07c74015.

The gate_relation follows the revised live gates in #183.

Evidence

• python3 -m unittest tests.test_incentive_monitor tests.test_incentive_forward: 34 OK.
• Full python3 -m unittest: 3,978 OK, 419 skipped.
• scripts/validate.py and evidence_manifest.py --check: passed.
• Live dry runs on 2026-09-24 (timing guard bypassed in the harness only; no order) ran end to end in about 5 s:
  • 1 then 5 names selected, with 4 then 20 matched controls;
  • 6,581 operating companies;
  • the engine's own mover_runner.py check returned valid with the pinned config sha256;
  • the scan hash matched the decision record.
• Monitor entitlements measured with a paper key of an Algo Trader Plus user:
  • data REST limit header 10,000/min;
  • OPRA wildcard at about 920 trades/s;
  • news wildcard connected;
  • the SIP stock stream is left to the engine (one connection per endpoint).

Limits

• The board weights are heuristic, fixed without outcome data.
• Paper fills are simulated.
• Licensed data stays private; nothing per-event is committed.

🤖 Generated with Claude Code

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.

[seathatflowsinourveins]

CI note (2026-09-25). validate-macos became a required check on main today (#226 and ruleset 23739774). This PR's latest CI run predates that change, so it has no validate-macos status, and GitHub shows the check as expected and waiting.

Nothing in the PR needs to change. The next push, or a close and reopen, runs the current adoption-bootstrap.yml, which reports validate-macos on every pull request.

🤖 Generated with Claude Code

[seathatflowsinourveins]

Trading lane, 2026-09-25: carrying this PR to merge.

Under the agreed lane split this PR is trading-lane work, and it has had no commits since 2026-09-24. The trading-lane coordinator is carrying it, per the user's approval of dormant-owner merges. The original owner can still take it back by replying here.

What changed:

• main (269602b) was merged in as 14bf5fc. The merge was clean and there was no force-push.
• The PR adds files only.
• The push re-runs CI on the current workflows, including the required validate-macos check.

Checks on the merged tree:

• tests.test_incentive_monitor tests.test_incentive_forward: 34 OK, 0 skipped, on both system Python and the pinned adaptive-paper runtime (Python 3.12.3, nautilus-trader 2.0.0rc5).
• scripts/validate.py, evidence_manifest.py --check and validate_convergence.py --all-recorded pass.

After the engine P0 release (#221, #242, #243):

• forward-protocol-v1.json pins only the two configs and the cost table. The engine files are hashed into each decision record, not pinned.
• All 8 ENGINE_FILES still exist on main, so a v1 decision would record the post-P0 engine hashes.
• No v1 order has been placed, and the paper keys currently return HTTP 401.
• Incentive monitor v2: status/LULD, corporate-action events, screeners, 20 s sweeps, option IV, FINRA short volume (board v2) #220 already plans a forward protocol v2 that pairs board v2 with the P0 engine, with its own pre-outcome review.

I'll merge once the required checks pass. #220, which is stacked on this PR, will then be retargeted to main.

🤖 Generated with Claude Code