Corrections log

Standing contact: corrections@thecompounding.company. Every material correction, suspension or removal is logged here with reasons.

A standing corrections contact is published, every correction is logged, and the log is public.

Corrections made during edition one, all self-reported:

• A scoring inconsistency in the source framework was found and fixed during development.
• Three anchor ambiguities surfaced after the first three companies and were resolved by written amendment before scoring continued.
• Three entity rulings were corrected mid-sector after verifying corporate structure against annual reports: the correct listed entity following a demerger; an overlap where a parent's results include separately scored subsidiaries; and a company recorded as an associate that was in fact a consolidated subsidiary.
• Reading the source instrument after thirty-seven files had been scored revealed three errors: a wrong band set in working notes, a ceiling that had been recorded rather than applied, and a confidence system computed on the wrong basis. The last changed the published grades of twenty-two companies.
• A first-pass confidence recompute was itself wrong for eight companies and was corrected on verification against the archived packs.
• A retail evidence passage was misread from a multi-column extraction and corrected on re-reading, which sharpened a cross-sector finding rather than weakening it.
• Absence scoring (Batch-2 item B2-1) was withdrawn for edition one after review showed it had been applied to six of fifty companies and interacts incorrectly with the edition-one denominator rule. Six scores were revised upward; no company's confidence grade or publication form changed. The rule is pre-registered for edition two.
• Six company files carried arithmetic errors in a dimension percentage or total (range −0.7 to +1.8 points); all were corrected on reconciliation to a locked calculation sheet. No band or grade changed.

An index that publishes its own corrections is worth more than one that never finds any. These are listed because the audit trail is the product.

Removal policy

The publisher will not remove an entry merely because its conclusion is disputed. Material factual errors will be corrected promptly. Publication may be suspended or removed where the underlying evidence cannot be verified, the wrong entity was assessed, procedural fairness materially failed, publication would violate law, or continued publication would create disproportionate and avoidable harm. Every material correction, suspension or removal will be logged with reasons.

Pre-publication corrections to published pages

30 September 2026 — index page, ordinal ranking added

The index page previously listed companies alphabetically within band and stated that no ordinal position was published. The forty-one companies carrying a published score are now ranked by that score, highest first, to one decimal, with equal scores sharing a rank and listed alphabetically; the nine band-only (Grade C) companies are not ranked and appear at the end. The methodology §6 listing rule and its change log were updated to match (v2.4 to v2.5). No score, band, grade, ceiling outcome or scoring rule was affected.

23 September 2026 — methodology page, §8 ceilings table

The ceilings table reported the Outcome, Context and Exposure ceilings as triggered twice each. The text beneath the table states, correctly, that these three ceilings cannot fire in edition one because every scored question carries a floor of 1, and no company file or page records any of them. The table figures were left over from a withdrawn draft and now read 0. The Security ceiling count of 10 is unchanged. No score, band, grade, ceiling outcome or scoring rule was affected. Page moved from v2.3 to v2.4.

9 September 2026 — methodology page, §9 and §13

The methodology page stated that a replication kit including blinded evidence packs was published under CC BY-SA, and listed that kit among the CC BY-SA material. No blinded evidence packs were built for edition one: sources were cited from the public record rather than assembled into per-company packs, so there was nothing to blind. The claim was removed and §9 now states what is published — the anchor set, the scoring rules, the sector rulings, and the source archive register on request — and why the packs do not exist. §13 was corrected to match. No score, band, grade, ceiling or scoring rule was affected. Page moved from v2.0 to v2.3.

9 September 2026 — sector rulings published

§5 promised that all sector rulings were published alongside the methodology; they had not yet been posted. The seven sector calibration notes, the anchor amendments and the v2.1 ratification and freeze record are now published, and the ruling text is unchanged. Status labels were corrected from "proposed" to "ratified": the notes were drafted on 30 August 2026, each carries on its face the instruction to ratify before its sector's first company file, and the draft labels were never updated.

Ratification was logged in the header of the following sector's note. That chain is uneven, and the rulings index now states exactly what it holds: a date for the BFSI and consumer-internet sets, ratification recorded without a date for healthcare and retail, and no separate ratification entry for manufacturing and telecom. Edition two will keep a single ratification register with a date against every ruling set.

Correction to this entry, 9 September 2026. As first published, this entry stated that each ruling set "was ratified on 30 August 2026" and that the chain merely lacked dates for some sets. That asserted more than the record contains: for manufacturing and telecom there is no separate ratification entry. The entry has been corrected and the overstatement is logged here rather than silently replaced.

9 September 2026 — company files, before publication

Four corrections were made to company files before any entry was published: a contradiction between the scored value and the amendment block in one file was resolved in favour of the amendment; the date of a source supporting one score was stated explicitly after review showed it predated the evidence window and rested on an in-window continuation; and two entries were re-sourced after cited pages became unavailable. Separately, a line indicating which additional evidence would raise which scores was removed from forty-three company files, and a reference to a paid assessment product was removed from all fifty files and from the methodology page. No score, band or grade changed.

Post-publication entries

None yet. Entries appear here once the Index is published.