Correct dated HHI threshold guidance for operators

[stranske]

Source: Issue #1023

Closes #1023

Automated Status Summary

Scope

docs/concentration_metrics.md:77 attributes the 0.18/0.25 concentration labels to DOJ conventions without a date. The current official DOJ HHI explainer describes a market above 1,800 as highly concentrated under the 2023 Merger Guidelines, so the 2,500 attribution is dated and the text mixes regimes. The document already calls these orientation heuristics, not policy limits. Preserve that distinction: this is a documentation correction, not a request to impose antitrust thresholds on counterparty risk.

Tasks

• Revise docs/concentration_metrics.md operator interpretation paragraph to label any historical threshold with its year or replace it with a dated current official reference.
• Keep docs/concentration_metrics.md explicit that merger-market heuristics do not set portfolio counterparty limits; point operators to the existing limit-monitoring guide.

Acceptance criteria

• Documentation verification: the changed paragraph links to the official DOJ HHI explainer, identifies the guideline vintage, and does not describe 2,500 as the undated current highly-concentrated threshold.
• Deliberate-break verification: temporarily restore the old undated 2,500 attribution in docs/concentration_metrics.md; the documented source-comparison checklist must fail; restore the correction and record the comparison in the PR.
• The documented fractional-to-10,000 conversion and actual calculation and configuration behavior remain unchanged.

[github-actions]

Workflow state fingerprint for Keepalive Loop Reporter. Do not edit.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Essentials

Run ID: 86a86e5d-2979-4cec-9b25-8fdc79253e61

📥 Commits

Reviewing files that changed from the base of the PR and between 6bee396 and 407fa21.

📒 Files selected for processing (1)

• docs/concentration_metrics.md

Included review availability: 0 reviews are currently available. Your included PR review attempts over the past 7 days set your current allowance at 1 review per hour.

---

📝 Walkthrough

Walkthrough

The documentation replaces undated HHI threshold guidance with a DOJ reference to the 2023 Merger Guidelines. It clarifies that the threshold is informational and directs operators to configured portfolio limits and breach-handling guidance.

Changes

HHI guidance correction

Layer / File(s)	Summary
Update HHI interpretation guidance docs/concentration_metrics.md	The operator guidance links the DOJ HHI explainer, references hhi > 0.18 under the 2023 Merger Guidelines, and separates market-concentration heuristics from portfolio limits.

Estimated code review effort: 1 (Trivial) | ~2 minutes

Severity of issue fixed: Low

Merge Risk: ⚪ Minimal · up to 407fa

The updated concentration guidance corrects the dated HHI reference and directs operators to existing portfolio-limit documentation without changing calculations, configuration, or runtime behavior.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly identifies the main change: correcting dated HHI threshold guidance for operators.
Linked Issues check	✅ Passed	The documentation changes satisfy issue #1023. They add the official DOJ HHI explainer and 2023 guideline reference, use the 1,800-point threshold, preserve the fractional-to-10,000 conversion, distin…
Out of Scope Changes check	✅ Passed	The changes are limited to the requested operator documentation. No unrelated workflow, source-code, calculation, configuration, or portfolio-threshold changes are present.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0…

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch codex/issue-1023-date-hhi-guidance

---

_{Comment @coderabbitai help to get the list of available commands.}

[agents-workflows-bot]

🤖 Keepalive Loop Status

PR #1043 | Agent: Codex | Iteration 0/12

Current State

Metric	Value
Iteration progress	[----------] 0/12
Action	wait (gate-not-success)
Disposition	skipped (failure)
Agent status	✅ ALL TASKS COMPLETE
Gate	unknown
Tasks	10/10 complete
Timeout	45 min (default)
Timeout usage	0m elapsed (2%, 45m remaining)
Keepalive	✅ enabled
Autofix	❌ disabled

🔍 Failure Classification

| Error type | infrastructure |
| Error category | unknown |
| Suggested recovery | Capture logs and context; retry once and escalate if the issue persists. |

[agents-workflows-bot]

Keepalive Work Log (click to expand)

#	Time (UTC)	Agent	Action	Result	Files	Tasks	Progress	Commit	Gate
0	2026-09-10 01:09:48	Codex	wait (gate-not-success)	skipped	—	0	10/10	—	—
1	2026-09-10 01:13:04	Claude	run (verify-acceptance)	success	—	0	5/5	—	success
1	2026-09-10 01:39:00	Codex	run (agent-run-failed)	failure	31 file(s)	0	5/5	—	success
1	2026-09-10 02:39:43	Codex	run (agent-run-failed)	failure	31 file(s)	0	5/5	—	success
1	2026-09-10 03:36:29	Codex	run (agent-run-failed-repeat)	failure	31 file(s)	0	5/5	—	success
1	2026-09-10 03:39:07	Codex	run (agent-run-failed-repeat) retry	failure	31 file(s)	0	5/5	—	success
1	2026-09-10 04:35:50	Codex	run (agent-run-failed-repeat)	failure	31 file(s)	0	5/5	—	success
1	2026-09-10 04:38:09	Codex	run (agent-run-failed-repeat) retry	failure	31 file(s)	0	5/5	—	success
1	2026-09-10 05:29:39	Codex	wait (gate-not-success)	skipped	—	0	5/5	—	—
1	2026-09-10 05:39:47	Codex	run (agent-run-failed-repeat)	failure	31 file(s)	0	5/5	—	success
1	2026-09-10 05:42:05	Codex	run (agent-run-failed-repeat) retry	failure	31 file(s)	0	5/5	—	success

[github-actions]

Workflow state fingerprint for Agents Gate Followups. Do not edit.

[github-actions]

Runner dispatch state for claude on PR #1043. Do not edit.

[stranske-keepalive]

🤖 Keepalive Loop Status

PR #1043 | Agent: Codex | Iteration 1/12

Current State

Metric	Value
Iteration progress	[#---------] 1/12
Action	run (agent-run-failed-repeat)
Agent status	✅ ALL TASKS COMPLETE
Gate	success
Tasks	5/5 complete
Timeout	45 min (default)
Timeout usage	2m elapsed (5%, 43m remaining)
Keepalive	✅ enabled
Autofix	❌ disabled

Last Codex Run

Result	Value
Status	❌ AGENT FAILED
Reason	agent-run-failed-repeat
Exit code	unknown
Failures	8/3 before pause

To retry immediately:

• Add the agent:retry label to this PR

Or wait for the next successful Gate run to automatically retry.

🔍 Failure Classification

| Error type | infrastructure |
| Error category | unknown |
| Suggested recovery | Capture logs and context; retry once and escalate if the issue persists. |

⚠️ Failure Tracking

| Consecutive failures | 8/3 |
| Reason | agent-run-failed |

🔁 Paused – Automation Recovery Required

The keepalive loop paused this execution strategy after repeated failures; ownership remains with automation.

To resume:

1. Route the failure to CI repair, retry/backoff, alternate-agent, review fallback, or issue decomposition
2. Record a concrete next action and responsible automation worker
3. Use needs-human only after an independent review proves a real authority boundary
4. Re-run Gate or apply the automation retry path

Or manually edit this comment to reset failure: {} in the state below.

[Copilot]

🟢 Approval recommended

The change is a scoped documentation correction that aligns the operator guidance with the cited official DOJ threshold and preserves the stated non-goals (no calculation/config behavior changes).

Pull request overview

Updates operator-facing HHI threshold guidance in docs/concentration_metrics.md to remove an undated 2,500-point “highly concentrated” attribution and instead cite the DOJ’s HHI explainer and the 2023 Merger Guidelines (page updated Jan 17, 2024), while preserving the existing distinction between merger-market heuristics and portfolio counterparty limits.

Changes:

• Replace the operator interpretation item with a dated DOJ source link and the correct “highly concentrated” threshold (HHI > 1,800 / hhi > 0.18 on the fractional scale).
• Clarify that merger-market concentration guidance is orientation-only for operators and does not define portfolio counterparty limits.
• Add pointers to existing Limit Monitoring documentation and config/limits.yml for configured exposure limits.

File summaries

File	Description
docs/concentration_metrics.md	Corrects and sources the operator HHI threshold guidance and reasserts the separation between merger-market heuristics and portfolio limit configuration.

Review details

• Files reviewed: 1/1 changed files
• Comments generated: 0
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[stranske]

Independent Orchestrator/Cursor advisory at exact head 407fa21cc83036b116e1c351e1ff41a051729119: approve; no P0/P1 blockers. It verified the one-paragraph diff, source #1023, official DOJ source snapshot, preserved 10,000 conversion, existing limit-monitoring link, complete criteria and corrected/old/restored source-comparison evidence (6 PASS / 5 FAIL + 1 PASS / 6 PASS). This is advisory evidence, not a substitute claim about CodeRabbit approval.

Reviewer noted that the moderate-concentration band is omitted. That does not violate #1023: the issue permits replacing the external analogy, requires correcting the highly concentrated citation, and leaves portfolio thresholds to configured limits. No broader band table is needed for this scoped correction.

Gate passed for the documentation scope; no runtime changes or runtime-test claims. Conservative review floor is 2026-09-10T01:16:33.315467+00:00, restarting on any new head. Keepalive/closer must recheck the unchanged head, expected/required checks and full active review threads after that floor before merge and verify:compare. No merge or auto-merge armed by the opener.

[github-actions]

Runner dispatch state for codex on PR #1043. Do not edit.

[stranske]

Exact-head check-presence disposition for 407fa21cc83036b116e1c351e1ff41a051729119 before merge. A fully paginated comparison against 12 recently merged Counter_Risk heads found eleven absent names. The prescribed root reporter is missing; it was not treated as PASS.

• The eight Python CI / ... children (Validate inputs, lint-format, lint-ruff, logs summary, python 3.12, python 3.13, select reusable CI scope, typecheck-mypy) are intentionally absent because current pr-00-gate.yml skips the reusable Python job on docs-only changes. This PR changes only docs/concentration_metrics.md. Gate run 34424299352 has successful path classification and gate-summary and an explicit Python CI SKIPPED record. This is documentation-scope validation, not a runtime-test PASS claim.
• emit-reference-run and conformance / Backplane run-contract conformance belong to backplane-conformance.yml, whose actual pull_request paths cover src, artifacts, scripts, docs/contracts and config/backplane_participants.json. The changed concentration guide matches none of them.
• verifier / Run post-merge verifier requires merge and the intended verifier label. The closer will apply verify:compare after merge.

Full check rollup contains 182 contexts; latest Gate / gate succeeds. Historical startup cancellations are superseded by successful Gate 34424299352; cancelled Autofix context/dispatch bookkeeping did not leave code work: source tasks are complete, exact-head Copilot recommends approval, and independent advisory 5611157678 records the source-comparison red/green evidence. Later runner failures report No output captured after acceptance verification completed; they are infrastructure bookkeeping rather than a product failure or human decision. No workflow, test, or required check was waived. Branch rules return []; normal product Gate still enforced. Final unchanged head and zero-active-thread recheck required, with the 01:16:33Z review floor already elapsed.

[github-actions]

Provider Comparison Report

Provider Summary

Provider	Model	Verdict	Confidence	Summary
openai	gpt-5.6-terra	PASS	87%	The operator guidance now links to the official DOJ HHI explainer, explicitly identifies the 2023 Merger Guidelines and January 2024 explainer update, and correctly states the current 1,800-point /...
anthropic	claude-sonnet-5	PASS	82%	The documentation change to docs/concentration_metrics.md correctly satisfies the primary acceptance criteria: it links to the official DOJ HHI explainer, cites the 2023 Merger Guidelines vintage...

📋 Full Provider Details (click to expand)

openai

• Model: gpt-5.6-terra
• Verdict: PASS
• Confidence: 87%
• Scores:
  • Correctness: 10.0/10
  • Completeness: 9.0/10
  • Quality: 9.0/10
  • Testing: 8.0/10
  • Risks: 9.0/10
• Summary: The operator guidance now links to the official DOJ HHI explainer, explicitly identifies the 2023 Merger Guidelines and January 2024 explainer update, and correctly states the current 1,800-point / 0.18 highly-concentrated reference without presenting 2,500 as an undated current threshold. It also preserves the required distinction that merger-market heuristics are informational only and directs operators to Limit Monitoring and config/limits.yml for actual portfolio controls. The fractional-scale conversion statement remains intact and no concentration calculation or configuration behavior was changed. The unrelated chat parsing hardening is well-scoped and improves robustness by rejecting NaN, infinities, overflowed numeric strings, and integers too large to convert to finite floats; added tests cover these cases. No material correctness, compatibility, security, or maintainability risk is apparent from the merged changes.

anthropic

• Model: claude-sonnet-5
• Verdict: PASS
• Confidence: 82%
• Scores:
  • Correctness: 9.0/10
  • Completeness: 8.0/10
  • Quality: 8.0/10
  • Testing: 8.0/10
  • Risks: 6.0/10
• Summary: The documentation change to docs/concentration_metrics.md correctly satisfies the primary acceptance criteria: it links to the official DOJ HHI explainer, cites the 2023 Merger Guidelines vintage, removes the undated 2,500/0.25 'highly concentrated' language, and explicitly redirects operators to the Limit Monitoring guide for actual portfolio limits — preserving the heuristic-vs-policy distinction requested. The fractional-to-10,000 conversion note and surrounding context remain otherwise intact. The PR also includes an unrelated but reasonable and well-tested hardening of _parse_float in chat/session.py to reject NaN/Infinity/overflow values, which is out of scope for this documentation-focused issue but does not conflict with or break the documented acceptance criteria. Overall the core documentation fix is correct, complete, and testable via manual inspection; the added code change is a minor scope-creep concern but does not undermine the PR's stated goals.
• Concerns:
  • The PR bundles an unrelated refactor of _parse_float in src/counter_risk/chat/session.py (adding NaN/Inf/OverflowError handling) that is outside the stated scope of a documentation-only correction. While well-tested, this introduces behavioral changes not covered by the PR's acceptance criteria or issue description, which could complicate review/audit trails.
  • The doc change removes the 0.25/2,500 'highly concentrated' heuristic entirely rather than providing a dated version of it; this satisfies the acceptance criterion (does not attribute 2,500 as an undated current threshold) but slightly narrows the original two-tier guidance (0.18 concentrated / 0.25 highly concentrated) to a single reference point (1,800/0.18) tied to the 2023 Merger Guidelines context.
  • No explicit mention in the diff of the 'deliberate-break verification' step being performed and recorded in the PR body (per acceptance criteria), though this may have been done outside the diff (e.g., in the PR description) and is not visible here.

Agreement

• Verdict: PASS (all providers)
• Correctness: scores within 1 point (avg 9.5/10, range 9.0-10.0)
• Completeness: scores within 1 point (avg 8.5/10, range 8.0-9.0)
• Quality: scores within 1 point (avg 8.5/10, range 8.0-9.0)
• Testing: scores within 1 point (avg 8.0/10, range 8.0-8.0)

Disagreement

Dimension	openai	anthropic
Risks	9.0/10	6.0/10

Unique Insights

• openai: The operator guidance now links to the official DOJ HHI explainer, explicitly identifies the 2023 Merger Guidelines and January 2024 explainer update, and correctly states the current 1,800-point / 0.18 highly-concentrated reference without presenting 2,500 as an undated current threshold. It als...
• anthropic: The PR bundles an unrelated refactor of _parse_float in src/counter_risk/chat/session.py (adding NaN/Inf/OverflowError handling) that is outside the stated scope of a documentation-only correction. While well-tested, this introduces behavioral changes not covered by the PR's acceptance criteria or issue description, which could complicate review/audit trails.; The doc change removes the 0.25/2,500 'highly concentrated' heuristic entirely rather than providing a dated version of it; this satisfies the acceptance criterion (does not attribute 2,500 as an undated current threshold) but slightly narrows the original two-tier guidance (0.18 concentrated / 0.25 highly concentrated) to a single reference point (1,800/0.18) tied to the 2023 Merger Guidelines context.; No explicit mention in the diff of the 'deliberate-break verification' step being performed and recorded in the PR body (per acceptance criteria), though this may have been done outside the diff (e.g., in the PR description) and is not visible here.

🔍 LangSmith Traces

• openai
• anthropic

[github-actions]

Workflow state fingerprint for Agents Verifier. Do not edit.

[stranske]

Verified completion through merged #1043. The durable provider comparison (#1043 (comment)) is PASS from OpenAI 87% and Anthropic 82%; the complete review-thread page is empty.

The actual PR diff/file list changes only one paragraph of docs/concentration_metrics.md. The report's chat-parser scope concern refers to earlier base work delivered separately by #1042, not a bundled #1043 change. Current main retains the dated 2023 guideline reference, the 1,800/0.18 conversion and explicit separation from configured portfolio controls. The source issue allows replacing the obsolete analogy; adding a broader band table is outside its acceptance criteria. The recorded source-comparison-results.json has corrected 6/6 PASS, old-text 5 FAIL/1 PASS, restored 6/6 PASS; exact-head advisory 5611157678 independently audited that evidence. No runtime behavior or portfolio threshold was changed.

All source criteria and report caveats are dispositioned, with no follow-up debt. Closing #1023 after durable verification.