Skip to content

History

Revisions

  • Follow the repository rename and the current roadmap Point every link at danielep71/K-SACCR instead of danielep71/VBA-SACCR-Toolkit: files, issues, releases, milestones, commits, image URLs, and the clone and cd lines in Getting Started. Bring the roadmap in line with the README on main: v1.7.0 is the standalone Python engine (#450-#459); the v1.8.0 to v1.10.0 rows and the note that v1.7.0 is unused are removed. Securities financing transactions are developed in K-SFTs; risk-weighted assets, capital and default-fund contributions to CCPs in K-Risk Engines; CVA risk in an engine of its own. Open Items and Limitations, Regulatory Basis and SA-CCR Overview say the same. Checked: no old repository name is left; every blob, tree and raw link resolves on main, anchors included.

    @danielep71 danielep71 committed Oct 27, 2026
    74e9663
  • Explain the missing v1.7.0 milestone in the roadmap The roadmap jumps from v1.6.0 to v1.8.0. A line under the table now says that v1.7.0 is not used: its CCP scope was delivered in v1.0.2 (#112) and default-fund contributions moved to v1.10.0 (#400), as the README roadmap says. Validated: git diff --check clean.

    @danielep71 danielep71 committed Oct 23, 2026
    6493e85
  • Add workbook protection and #357 to the roadmap The v1.1.0 row now covers workbook protection (#420, #421, #422) alongside CSV import and staging, and the v1.2.0 row lists the Restore Parameter Defaults command (#357), matching the milestones and the README roadmap. Validated: git diff --check clean; every issue of the v1.1.0 and v1.2.0 milestones appears in its row.

    @danielep71 danielep71 committed Oct 23, 2026
    3ba5aa3
  • Keep the distinct-author condition for independent values The independent class said "an external implementation written outside the project", which dropped the condition that the implementation not be written by the author of the code under test. It now says "an external implementation by another author", on Validation Approach and Regulatory Sources. Validated: git diff --check clean; no "owner" or "maintainer" left in the wiki.

    @danielep71 danielep71 committed Oct 23, 2026
    c2dd803
  • Update branch status and drop third-person references to the author Open Items and Limitations now lists release/1.0.0, release/1.0.1 and release/1.0.2 as closed, after the retirement of the historical release branches. Replace "the owner" and "the maintainer" with neutral wording on ten pages: Home, Getting Started, Test Catalogue, Workbook Guide, Validation Approach, Regulatory Sources, Regulatory Basis, Worked Examples and Replacement Cost and Collateral. Measurements, runs and screenshots are described without naming who made them; the evidence-policy text says the project has no separate reviewer and that values derived within the project stay illustrative. The meaning of the evidence classes is unchanged. Validated: no "owner" or "maintainer" left in the wiki; git diff --check clean.

    @danielep71 danielep71 committed Oct 23, 2026
    d1d0728
  • Update the wiki for release 1.0.2 Inputs, Margining and MPOR, Outputs and Statuses, Troubleshooting and Getting Started describe the CCP columns R:S, roles CCP and MULTI, the settled-to-market columns AB:AC with Comment in AD, the CCP MPOR floor and retained-VM horizon, the new Checks INFO lines and the light-red highlight, and moving inputs from 1.0.1. Test Catalogue, Validation Approach and Open Items give the 1.0.2 counts (288 cases, 1,283 checks), the certification at cfb88e5 with its one deviation, the evidence rule without independent review, and the remaining limitations. The regulation pages, Home and the footer name release 1.0.2.

    @danielep71 danielep71 committed Oct 23, 2026
    be8f921
  • Keep one Basel vs CRR page Moves "Why two regimes" into Regulation/Basel-vs-CRR.md and removes the duplicate page at the wiki root, so the README link opens a single page. The sheet comparison now counts 28 areas, matching the Basel_vs_CRR rows added in release 1.0.2.

    @danielep71 danielep71 committed Oct 20, 2026
    09a7f74
  • Explain why the toolkit has two regimes Adds "Why two regimes" to the Basel vs CRR page: CRR is the default and the binding rule set for EU institutions; BCBS does not replace it but shows the effect of the EU-specific deviations, traces figures to CRE52, and serves Basel monitoring exercises, peer comparison and group entities outside the EU. BCBS results are not regulatory figures.

    @danielep71 danielep71 committed Oct 20, 2026
    0c519e8
  • docs(wiki): clarify worked example units precision and evidence

    @danielep71 danielep71 committed Oct 18, 2026
    229d7e4
  • docs(wiki): distinguish Basel CRR rules inputs and interpretations

    @danielep71 danielep71 committed Oct 18, 2026
    e9e45bb
  • docs(wiki): clarify PFE inputs aggregation and regulatory limits

    @danielep71 danielep71 committed Oct 18, 2026
    dd5ba24
  • docs(wiki): fix math spacing in GitHub rendering

    @danielep71 danielep71 committed Oct 18, 2026
    269991d
  • docs(wiki): clarify margining and MPOR rules and limitations

    @danielep71 danielep71 committed Oct 18, 2026
    2bd3cd1
  • docs(wiki): clarify replacement cost and collateral treatment

    @danielep71 danielep71 committed Oct 18, 2026
    9c7d88e
  • docs(wiki): correct and complete regulatory sources

    @danielep71 danielep71 committed Oct 18, 2026
    33ba1ea
  • docs(wiki): correct regulatory basis and coverage limits

    @danielep71 danielep71 committed Oct 18, 2026
    f332297
  • Copy BCBS 279 and register the CRR CCP extract BCBS 279 (March 2014, revised April 2014) is now kept with the other source copies, with its SHA-256 hash: the BIS terms allow non-commercial redistribution, which also covers the Basel chapters already copied. The page no longer says BCBS 279 is cited but not copied, and states that the copies stay under their publishers' terms. CRR-CCP (Articles 4(1)(88)-(91) and 300-311 of the consolidated CRR) joins the source register and the hash table.

    @danielep71 danielep71 committed Oct 18, 2026
    946419c
  • docs(wiki): correct and complete SA-CCR overview Clarify output-floor alpha, aggregation, adjusted notionals, netting and method eligibility. Refine scope and validation wording while preserving glyphs, formatting, badges and images.

    @danielep71 danielep71 committed Oct 18, 2026
    d99d0a2
  • Give the demonstration portfolio's current input fingerprint How It Works (RunSACCR_Silent line) and Outputs and Statuses (Results C4 example) show inputs 7B68D18F, the value of the 1.0.1 release workbook; 74A6C31F was recorded before later changes to the template's inputs.

    @danielep71 danielep71 committed Oct 18, 2026
    2292b34
  • Make the source descriptions exact Regulatory Sources: the consolidated CRR text is a documentation tool with no legal effect; K-SACCR implements Articles 274-280f, and 273a, 281 and 282 are context only. BCBS 279 is available from the BIS publication page and is cited, not copied, because its notice reserves all rights. Only published and independent values qualify as numerical-validation references; CRR provisions were checked, CRR-specific numerical validation remains incomplete. Regulatory Sources, Margining and MPOR, Basel vs CRR: the client's 10-day MPOR floor under Basel depends on the conditions of CRE54.15 (CRE54.14), CRE54.16 changes only the risk weight to 4%, CRE54.17 makes a non-qualifying arrangement a bilateral trade, and CRE54.8(2)'s 10-day minimum is stated for OTC derivatives. Matches the repository after #375.

    @danielep71 danielep71 committed Oct 18, 2026
    ec08026
  • Describe release 1.0.1 Status, badges, footer and download steps name 1.0.1 (tag v1.0.1, 2026-10-18); the 1.0.1 certification is added beside the 1.0.0 record; suite counts are 230 cases and 1,033 checks; the active branch is release/1.0.2; CCPs (#112) are planned in v1.0.2 instead of v1.7.0.

    @danielep71 danielep71 committed Oct 18, 2026
    3c62050
  • SA-CCR explainers to the overview, RC, and PFE wiki pages

    @danielep71 danielep71 committed Oct 18, 2026
    1e387e3
  • Describe the inputs-only C5 status and EDITED Results C5 now reads "Inputs unchanged since the last run (inputs ...); outputs are checked when evidence is exported" when the inputs match the last run, and ResultsStatus() gains EDITED for an output cell changed by hand after the run, which the evidence export refuses. The pages show the new text with the 1.0.0 wording beside it, and mark EDITED as from 1.0.1, so they stay correct for the released workbook: - Outputs-and-Statuses: C5 table with the new text and an EDITED row; a paragraph on why C5 checks the inputs only. - Reconciliation-and-Evidence: refusal table names STALE or NONE, and EDITED for an edited output cell. - Workbook-Guide: C5 row lists the new text. - Getting-Started: expected C5 text after the first run. - How-It-Works: ResultsStatus() returns CURRENT, EDITED, STALE or NONE.

    @danielep71 danielep71 committed Oct 17, 2026
    aea9ed0
  • Update Troubleshooting.md

    @danielep71 danielep71 committed Oct 17, 2026
    c93df0d
  • Show where the CRR uses SA-CCR New section on SA-CCR Overview with the diagram from docs/assets (main): counterparty credit risk, CCP exposures, large exposures and CVA risk take the risk-based exposure value; the leverage ratio (Art. 429c) and the output floor (Art. 465(4)) recalculate SA-CCR with their own adjustments. A table gives the articles, dates, treatment and the K-SACCR milestone or scope of each.

    @danielep71 danielep71 committed Oct 17, 2026
    1e8a223
  • Align the wiki with release 1.0.0 Release 1.0.0 is published: status lines, footer and Home name the tag v1.0.0, its assets and the active branch release/1.0.1. Getting Started explains the released workbook (hash check, then unblock). The pages that quoted the template and messages at 0a6138c were checked against v1.0.0 and corrected: Params citations, output and reconciliation headers, C4/C5 texts, withheld netting sets split by regime, the README sheet reference card, the Basel_vs_CRR colours and the TestCatalogue source labels. Open Items and Validation Approach record the 1.0.0 certification and its declared deviation.

    @danielep71 danielep71 committed Oct 17, 2026
    51a062a
  • Update Validation-Approach.md

    @danielep71 danielep71 committed Oct 17, 2026
    410982e
  • Remove decision numbers and dates from the regulation and validation pages The methodology no longer records internal decision numbers or dates; settled points are stated under "Assumptions and scope". These pages now match: decision IDs and dates are dropped from Regulatory-Basis (the settled-points table and the parameter notes), Margining-and-MPOR, SA-CCR-Overview and Validation-Approach. The issue links stay. The link to "methodology › decision 8" now points to "Assumptions and scope".

    @danielep71 danielep71 committed Oct 17, 2026
    0928f1a
  • Update the used-range timing to the current benchmark (row 105,005, 1.79 s).

    @danielep71 danielep71 committed Oct 17, 2026
    47678f5
  • Report the 500,000-trade benchmark and say "No real portfolio data" The owner's benchmark of 2026-10-17 (#291) measured synthetic CRR portfolios on the supported 64-bit Windows host: 1,000 legs in 0.31 s, 10,000 in 1.62 s, 50,000 in 8.50 s and 500,000 legs in 1,000 netting sets in 113.93 s, every leg used and every netting set VALID. - Getting Started: new Portfolio size section with the table, no fixed trade limit, the 5,000-row validation range of Trades and the cost of an inflated used range. - Open Items and Limitations: the size measurement of #46 marked done; the limitation now states the measured size. - Home: Scale badge and the result in the introduction. - Footer and Regulatory Basis: "No real portfolio data" replaces "Synthetic data only", as in the repository (#293).

    @danielep71 danielep71 committed Oct 17, 2026
    df22009