[genedan]

Summary of Changes

Adds notebook to verify that the HurlimannCLR estimator, introduced in #1553, reproduces the values in the paper.

Related GitHub Issue(s)

#1502

AI/LLM Usage

Claude Opus 5.5

Additional Context for Reviewers

Submitter's Checklist

• I have reviewed and am adhering to the standards outlined in the project Governing Doc.
• The PR subject title summarizes the changes, with one proper prefix ([FIX], [FEAT], [DOCS], [TST], [CHORE], or [BRK]).
• I am a human (not a bot), and this PR form is written by a human.

Reviewer's Checklist

• The implementation addresses the associated issue(s).
• The implementation is appropriate, maintainable, and follows ARCHITECTURE.md.
• PR subject title has the proper prefix and the subject is appropriate.
• Relevant issue(s) are linked.
• AI/LLM usage is disclosed and appropriate.
• Documentation and tests are appropriate.
• CI tests passed, or any failures are acceptable.
• Leave a comment with the final recommendation (e.g. approve as is, request a secondary review, or flag an area for more review).

---

Note

Low Risk
Documentation-only notebook with assertions against published examples; does not modify reserving logic or runtime behavior.

Overview
Adds docs/user_guide/hurlimann_verification.ipynb, a user-guide notebook that checks cl.HurlimannCLR against the numeric tables in Hürlimann (Astin Bulletin 39(1)), for issue #1502.

The notebook rebuilds the paper’s paid triangles (incremental and cumulative), runs the five credibility method variants with exposure via sample_weight, and asserts outputs against published values for parameters (p_, m_, credibility_), IBNR/ultimates, and MSE ratios. Where the CAS Exam 7 errata fixes misprinted ultimates/MSE columns, the expected values use the corrected tables. For Table 7.12 (A.M. Best comparison), it documents that exposures are least-squares reconstructed because the paper omits them, and still checks all method factors to three decimals.

Low documentation-only change; no library API or estimator behavior changes.

^{Reviewed by Cursor Bugbot for commit 381fd8d. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.46%. Comparing base (c56cdbd) to head (381fd8d).

Additional details and impacted files

@@                 Coverage Diff                  @@
##           hurlimann-estimator    #1554   +/-   ##
====================================================
  Coverage                94.46%   94.46%           
====================================================
  Files                      100      100           
  Lines                     6307     6307           
  Branches                   803      803           
====================================================
  Hits                      5958     5958           
  Misses                     225      225           
  Partials                   124      124           

Flag	Coverage Δ
unittests	94.46% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

Pyright Type Completeness

View the full pyright --verifytypes output for this commit

Project (full chainladder package, at this PR's head): 16.1% of exported symbols fully typed (264 / 1636)

	Known	Ambiguous	Unknown	Total
Project (head)	264	105	1267	1636

Other symbols referenced but not exported by chainladder: 24

	Known	Ambiguous	Unknown	Total
Other (head)	11	1	12	24

Symbols without documentation:

• Functions without docstring: 347
• Functions without default param: 0
• Classes without docstring: 10

Patch (exported symbols added or changed by this PR): no exported symbol type-completeness changes detected.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 381fd8d. Configure here.}

[cursor]

Verification helper uses default rtol

Medium Severity

close calls np.allclose with only atol, so the default rtol=1e-5 is still applied. For the large whole-number totals in Tables 7.8 and 7.9 that intended tolerance of 1 becomes tens or hundreds of units, so the notebook can accept values that do not match the paper’s rounding.

 

^{Reviewed by Cursor Bugbot for commit 381fd8d. Configure here.}