*English edition (en-US), September 29, 2026. This English text is the reference version. It replaces the Turkish text of July 20, 2026 as the reference text; the Turkish text stays published for the record, and its Clause 7.6 no longer applies.*

# CARBONBENCH BADGE PROTOCOL (CB-B) v1.0
## Definition, Criteria Formulas, and Publication Procedure for Periodic Performance Badges

| | |
|---|---|
| **Document code** | CB-B v1.0 |
| **Version** | 1.0 (first publication) |
| **Date** | July 20, 2026 |
| **Status** | IN FORCE — a derivative document of CB-P v1.0-draft; passes to the oversight of the Methodology Board together with the final approval of CB-P |
| **Parent documents** | CB-P v1.0 (docs/protocol-v1.md) · Whitepaper v1.0 (internal document, not published) · Model Card v1.1 (internal document, not published) |
| **Copyright and license** | © 2026 Carbonbench. CC BY-ND 4.0 — may be reproduced with attribution and without modification. |

---

## 0. Table of Contents

1. Purpose and scope
2. Normative references
3. Badge types and criteria formulas (CB-B.1 Sector Leader · CB-B.2 Most Transparent · CB-B.3 Most Improved · CB-B.4 In-Band Accuracy)
4. Periodicity
5. Independence and conflict of interest
6. Sector annex framework
7. Limitations and disclaimer
Appendix A — Badge criterion ↔ normative reference mapping table (normative)
Appendix B — Terms (informative)
Appendix C — Open items (v1.1 candidates, informative)

---

## 1. Purpose and scope

**1.1** This Protocol governs the definition, criteria formulas, periodicity rules, and independence safeguards of the **periodic performance badges** published on the Carbonbench platform, and the production framework for the sector annexes.

**1.2 Founding principle — a badge is not a new measurement.** Each badge **is a periodic derivative** (SHALL) of three metrics already calculated and sealed within the scope of CB-P v1.0 — the **Accuracy Estimate** (CB-P §9.4.1), **Transparency Score v0.2** (`config/scoring.json` · `transparency_accuracy`) and **Target Quality v2** (`config/scoring.json` · `target_quality_v2`). Badge computation collects no new data, changes no score, and regenerates no seal; it only runs the sealed scores in the period snapshot through the closed-form criteria in this document. CB-P's reproducibility constraint (CB-P §2.5) extends to badges unchanged: with the same period snapshot and the same criteria version, any third party arrives at the same badge list (SHALL).

**1.3 Areas not governed by this Protocol:**
a) The calculation of the scores themselves (CB-P layers K0–K4);
b) The embedded profile badge (SVG generation by `pipeline/rozet.py`) — it is the **visual carrier** of the current score, not a periodic award; the subject of this Protocol is solely periodic award badges;
c) Verification of physical reality (CB-P §2.3-a, Principle I3): a badge documents the relative superiority of a disclosure's statistical plausibility; it does not guarantee that actual emissions are low or accurate.

**1.4** In this document, **SHALL** expresses a mandatory requirement, **SHALL NOT** an absolute prohibition, **SHOULD** a strong recommendation, and **MAY** a permission (CB-P §2.4 convention).

**1.5 Sector unit.** The unit of competition for badges is the **subsector**. In v1.0, subsector = the 17 sector groups of `pipeline/scoring.py` SECTOR_MAP; when a finer taxonomy (sector taxonomy v2) is published, the subsector definition passes to it by a version note, and period results prior to the transfer date are NOT RECALCULATED (SHALL NOT — the projection onto badges of the CB-P §8.6.1 retrospective stability principle).

---

## 2. Normative references

The following documents are indispensable for the application of this Protocol to the extent that they are referenced. For each reference, **where** it is used in the badge criteria is stated explicitly.

**2.1 GHG Protocol Corporate Accounting and Reporting Standard** — WRI/WBCSD, Revised Edition (2004; together with the Scope 2 Guidance 2015 amendment).
*Use:* The Scope 1/2/3 definitions and the organizational boundary concepts are taken from this standard; the "Scope 1+2" magnitude in all badge criteria rests on these definitions. The exclusion of biogenic CO₂ from Scope 1 (CB-P Appendix C.3) is also a requirement of this standard.

**2.2 GHG Protocol Scope 2 Guidance** (WRI, 2015).
*Use:* In the In-Band Accuracy badge (Clause 3.4), the band comparison is made only against the **location-based** figure (CB-P §6.2.7); a lack of dual reporting is a Transparency internal-consistency signal (CB-P Appendix C.2) and indirectly affects the Most Transparent badge.

**2.3 ISO 14064-1:2018** (organization-level GHG inventories) and **ISO 14064-3:2019** (verification and validation of GHG statements).
*Use:* The terminology of "inventory", "disclosure" and "assurance"; the assurance-class signals of the Transparency Score (the reasonable/limited assurance distinction) are consistent with the conceptual framework of this family. Badges are NOT an ISO 14064-3 verification opinion and cannot be presented as such (SHALL NOT; see Clause 7).

**2.4 CSRD — Directive (EU) 2022/2464** (December 14, 2022, corporate sustainability reporting) and **ESRS — Commission Delegated Regulation (EU) 2023/2772** (July 31, 2023), in particular **ESRS E1** (Climate Change).
*Use:* ESRS **E1-6** (the obligation to disclose gross Scope 1/2/3 and total GHG emissions, including intensity metrics) is the regulatory framework for the sector intensity comparison: the in-band disclosure requirement in the Sector Leader badge and the intensity medians in the sector annexes (Clause 6) lie on the same axis as the comparability objective envisaged by E1-6. ESRS **E1-4** (targets) is the regulatory counterpart of the Target Quality component. The coverage base of the European universe (which companies' reports can be expected) is read from the CSRD scope rules.

**2.5 SEC Climate Disclosure Rule** — "The Enhancement and Standardization of Climate-Related Disclosures for Investors", 17 CFR Parts 210, 229, 232, 239, 249 (Final Rule, March 2024; Release No. 33-11275 / 34-99678).
*Use:* Reference for the Scope 1/2 disclosure framework of large reporters in the US universe.
**HONEST NOTE (SHALL — repeated in every period report):** Implementation of the rule is STAYED due to litigation — the SEC stayed its effectiveness on its own initiative in April 2024 and in 2025 withdrew from defending the rule in court. This Protocol cites the SEC rule NOT as an *obligation in force* but as a *framework reference* for US disclosure practice; the badge assessment of US companies rests entirely on voluntary/publicly available disclosures.

**2.6 California SB 253 — Climate Corporate Data Accountability Act** (2023; with the SB 219/2024 amendments) and **SB 261** (climate-related financial risk report).
*Use:* SB 253 (a reporting obligation, for companies doing business in California with revenue >USD 1 billion, for Scope 1/2 — from 2026 — and Scope 3 — from 2027) and SB 261 (>USD 500 million, a climate risk report every two years) are the de facto driver of reporting in the US universe while the SEC rule is stayed; they are cited in the rationale footnote for the US coverage base.

**2.7 SBTi Corporate Net-Zero Standard** (Science Based Targets initiative).
*Use:* The SBTi component of Target Quality v2 (20/100) and the 1.5°C pace component (15/100; sectoral pathway cross-verification CB-P Appendix C.5); hence the TargetQuality term of the Sector Leader composite. The official SBTi list is the sole source of verification; a "science-based" self-declaration does not on its own generate points.

**2.8 IFRS S2 Climate-related Disclosures** (ISSB, June 2023).
*Use:* Global disclosure baseline; in the Türkiye universe, the coverage rules of TSRS 2 (the local adaptation of IFRS S2, KGK) rest on this standard. The TSRS/KGK reports that are the input to the badge metrics are produced within this framework. In its industry-based disclosure guidance, IFRS S2 refers to **SASB-based industry metrics** (2.9); the sector-annex architecture of this Protocol follows the same chain.

**2.9 SASB Standards** — Sustainability Accounting Standards Board sector standards (77 industries; today maintained under the ISSB/IFRS Foundation umbrella).
*Use:* The sectoral metric selection of the sector annexes (Clause 6) is aligned with the relevant GHG/energy metrics of the SASB industry standards; in the transition of the subsector definition to a finer taxonomy (Clause 1.5), the **SICS (SASB Sustainable Industry Classification System) ↔ NACE mapping** is the target classification bridge (for consistency with the NACE intensity bands of CB-P §5.2.1). In choosing the physical intensity denominator (tCO₂e/ton of product — Clause 6.2-c), the activity metric definition of the relevant SASB industry standard is taken as the basis.

**2.10 TCFD — Recommendations of the Task Force on Climate-related Financial Disclosures** (2017; monitoring responsibility transferred to the IFRS Foundation in 2024 — historical reference).
*Use:* The "target + base year + scope breadth" structure of the Target Quality component is a legacy of the TCFD Metrics & Targets pillar; the disclosure structure in pre-2023 reports (the historical layer of the universe) is read through this framework. New criteria are bound not to TCFD but to its successor, IFRS S2.

**2.11 GRI 305: Emissions** (GRI Standards, 2016).
*Use:* Most reports in the universe present their emissions disclosure in the GRI 305-1/2/3 (Scope 1/2/3) and 305-4 (intensity) structure; the extraction templates and the breakdown/intensity signals of the Transparency Score are consistent with this disclosure practice. (Operational note: confusing GRI disclosure codes with table values is a known reading trap and is filtered out deterministically.)

**2.12 CDP Climate Change Questionnaire**.
*Use:* It is NOT a binding criterion input (questionnaire responses are partly paid/access-restricted, and the platform uses only publicly available sources); if the company's publicly available CDP response/report exists, it may enter the evidence set as evidence of the cross-consistency of scope and target disclosures (MAY) and feeds the Transparency signals. The CDP score does not enter the badge criteria through any channel (SHALL NOT — separation of the behavior score from the accuracy layer).

**2.13 Internal normative references:** CB-P v1.0 (score formulas §9.4.1, band §9.5.1, caps §5.3.1 and Appendix C.1, dispute §11.5, retrospective stability §8.6.1) · Model Card v1.1 (version freeze calendar §5) · Reading Error Matrix (docs/error-matrix.md) · `config/scoring.json` v0.2 · `config/fiziksel_yogunluk.json` (fiz_yog_v1).

---

## 3. Badge types and criteria formulas

**3.0 Common eligibility prerequisites (all badges; SHALL):**
a) The record shall be **active** (not superseded) and shall have been sealed before the period cut-off (Clause 4.3);
b) The company's MOST RECENT record within the period is taken as the basis; one candidacy per company;
c) The record SHALL NOT be in "reading under verification" status (CB-P reading policy: while there is suspicion of our own reading error, the company is neither penalized nor rewarded);
d) A record with an active **score cap** on the Accuracy Estimate (the K0 intensity cap, the `beyan_tutarsizligi_agir` cap — CB-P §5.3.1, Appendix C.1) cannot be a candidate for any badge;
e) The badge is awarded not to the record but to the **company + period**, and is published with the period label (e.g., "Sector Leader — Energy · 2026-Q4").

### 3.1 CB-B.1 — Sector Leader

**3.1.1 Formula (SHALL):** For each eligible company i within subsector s, the composite score:

> **R_i = 0.5 · AccuracyEstimate_i + 0.3 · TransparencyScore_i + 0.2 · TargetQuality_i**

All three metrics are on the [0,100] scale; R_i ∈ [0,100]. The company with the highest R_i value in the subsector receives the badge.

**3.1.2 Additional conditions (SHALL):**
a) **min n = 8:** In the period snapshot, the subsector shall have at least 8 eligible companies (meeting the 3.0 prerequisites); otherwise, Sector Leader is NOT AWARDED in that subsector for that period (a "competition not opened" note is published).
b) **In-band disclosure requirement:** The candidate's disclosed Scope 1+2 total shall lie within its own 95% confidence band (CB-P §9.5.1). A company with an out-of-band disclosure cannot receive the badge even if it ranks first by R_i; the badge passes to the next eligible candidate and the situation is published with its rationale.
c) **Tie-break order:** (1) higher Accuracy Estimate → (2) narrower band width (smaller σ_ln,T) → (3) earlier seal date.

**3.1.3 Threshold rationale (informative):**
- *Weights (0.5/0.3/0.2):* The platform's primary claim is to be the accuracy layer; an absolute-majority weight for the Accuracy Estimate is the arithmetic counterpart of this claim. Transparency comes second because accuracy depends on evidence, and evidence depends on transparency; Target Quality, as the only forward-looking component, carries the lowest weight (the quality of past disclosure is more testable than a promise about the future).
- *min n = 8:* The within-sector percentile threshold for score production is n≥5 (`config/scoring.json` · `min_sector_n`); because awarding produces a ranking, a more conservative floor is required. At n=8, first place rests on at least 7 comparison observations, and the risk of a single outlier's "automatic championship" drops to an acceptable level.

**3.1.4 Sensitivity note (SHALL):** In every period report, whether the winner changes when the weight vector is perturbed by ±0.05 (the 0.55/0.25/0.2 and 0.45/0.35/0.2 variants) is published per subsector. If the winner changes under these perturbations, the badge is still awarded by the formula value, but a "sensitive first place" note is added — the fact that the ranking is data-driven as much as it is a matter of formula choice is not hidden from the reader.

### 3.2 CB-B.2 — Most Transparent

**3.2.1 Formula (SHALL):** The highest **Transparency Score v0.2** in the subsector (6 components: EF transparency 27 · GWP recency 13 · methodology 18 · internal consistency 14 · Scope 3 disclosure 18 · value chain disclosure 10).

**3.2.2 Additional conditions (SHALL):** min n = 8 (same as 3.1.2-a) · absolute floor **TransparencyScore ≥ 70** (relative first place is not enough; the least bad company in a weak subsector cannot be declared "Most Transparent") · tie-break: (1) value chain component score → (2) Accuracy Estimate → (3) earlier seal date.

**3.2.3 Rationale (informative):** The floor of 70 is the lower bound of the "moderate" band (70–84) of the site's score tiers: the badge rewards only a level of transparency that is also defensible in absolute terms.

### 3.3 CB-B.3 — Most Improved (momentum)

**3.3.1 Formula (SHALL):** For the company's Accuracy Estimate values in two consecutive period snapshots:

> **Δ_i = AccuracyEstimate_i(t) − AccuracyEstimate_i(t−1)**, condition: **Δ_i ≥ +5** points.

The highest Δ_i in the subsector receives the badge; if no one clears the +5 threshold, the badge is NOT AWARDED for that period.

**3.3.2 Two-period data requirement (SHALL):** The company shall have an eligible (3.0) score in both the t and the t−1 snapshots. A company scored for the first time cannot be a candidate.

**3.3.3 Netting out the formula effect (SHALL):** If the formula/table version changed between the two snapshots (Model Card §5 windows), the "reason for change" distinction in the score history (`skor_gecmisi`) is used: a change **caused by the formula version** cannot be counted as momentum (SHALL NOT); Δ_i is calculated only on snapshot pairs that are comparable under the same version, or on the component whose reason for change is "new evidence/new report". Past scores are not recalculated under any circumstances (CB-P §8.6.1).

**3.3.4 Rationale (informative):** The +5 threshold was chosen so as to stay above the typical noise, on the score plane, of the reading variability measured in the confirmation round (error matrix: 11.5% discrepancy rate among adjudicated records — lower bound) [Editor's note, September 29, 2026: 11.5% is the value measured on July 14, 2026; the current value is published in docs/error-matrix.md.]: what is rewarded is not small fluctuations but the submission of new evidence (EF breakdown, assurance upgrade, scope expansion). This is also incentive design: the reward side of CB-P's principle that "heavy data requests are a sales funnel".

### 3.4 CB-B.4 — In-Band Accuracy

**3.4.1 Criterion (SHALL):** This badge is NON-COMPETITIVE — every company that clears the threshold receives it:
a) The disclosed Scope 1+2 total is within the 95% confidence band (CB-P §9.5.1);
b) No active score cap (3.0-d);
c) No open gross-error flag (unit/scale/scope-copy) processed into the `q̂_kaba` channel during the period;
d) Accuracy Estimate ≥ 70.

**3.4.2 Rationale (informative):** Being in-band is not sufficient on its own, because a very wide band "encloses" every disclosure; condition (d) restricts this to the case where the band is narrow — that is, where the evidence is strong. The badge is the sealed counterpart of the statement "your disclosure is consistent with your own uncertainty budget"; it is a contribution, from the data-quality side, to the comparability objective of ESRS E1-6.

---

## 4. Periodicity

**4.1 Standard period (SHALL):** Calendar quarter (Q1: Jan–Mar, Q2: Apr–Jun, Q3: Jul–Sep, Q4: Oct–Dec). Calculation + publication is completed within **7 days** following the close of the quarter; badges are published with the period label (e.g., "2026-Q4").

**4.2 Launch-year exception — "Launch Edition" (MAY):** In the platform's launch year, badges may be awarded on a MONTHLY period; these badges permanently carry the **"Launch Edition"** label (SHALL) and are counted separately from the quarterly series. The exception expires automatically at the end of the launch year; extending it requires a Methodology Board decision.

**4.3 Data cut-off rule (SHALL):** Period snapshot = active records sealed as of 23:59:59 UTC on the last calendar day of the period. A record/correction that arrives after the cut-off moment — even if the calculation has not yet run within the 7-day window — CANNOT ENTER that period's snapshot (SHALL NOT); it affects the next period. The snapshot summary at the cut-off moment (record count, summary of the record-hash list) is sealed together with the period report; anyone can reconstruct the snapshot.

**4.4 Formula version freeze (SHALL):** The badge criteria version (this document) and the input score formula versions are aligned with the Model Card §5 freeze calendar (change windows: **January and July**). **Neither the score formula nor the badge criteria can change within a period** (SHALL NOT); window changes take effect only from the snapshot of the following period. Every period report carries, in its imprint, the criteria + formula + table versions used.

**4.5 Persistence (SHALL):** An awarded badge remains in the historical record with its period label; failing to win it in subsequent periods does not revoke the badge. Revocation occurs ONLY as the result of a Clause 5.3 dispute/correction, and the old record is not deleted but moved to "revoked/corrected" status (CB-P §11.5 pattern).

---

## 5. Independence and conflict of interest

**5.1 Free of charge and not purchasable (SHALL / SHALL NOT):** Badges are FREE. A badge, ranking position, or candidacy cannot be awarded, expedited, or influenced in exchange for any fee, subscription, partnership, or commercial relationship (SHALL NOT). Badge computation is INDEPENDENT of whether the company has claimed its profile: an unclaimed company also wins; a claim only facilitates notifications and access to the embedded badge file.

**5.2 Explicit distinction from pay-to-play models (informative, SHALL in publication texts):** In common software/service badge marketplaces (G2 and similar "badge" ecosystems), badges are typically intertwined with platform membership, review campaigns, or paid profile management, and the criteria formulas are closed to the public. CB-B is built on the opposite basis: (a) the criteria formulas are published in closed form in this document; (b) the input scores are sealed and reproducible by third parties; (c) money is not a variable at any stage; (d) the list of winners is published together with the full sector snapshot, in which the losers can also be seen. Every badge page carries an independence statement summarizing these four points.

**5.3 Dispute process (SHALL):** A **7-day dispute window** is open following the publication of the period results. Grounds for dispute are limited to CB-P §11.5 (input-hash mismatch · snapshot/criteria application error · incomplete assessment of the evidence set); a dispute against the criterion itself is directed to the Methodology Board consultation channel. When a dispute is upheld, the corrected result is sealed as a NEW version of the period report; the old version remains accessible with a "corrected" mark and is not deleted (SHALL NOT). The general dispute SLA for score records (10 business days, Model Card §6) applies separately and at all times.

**5.4 Margin-of-error statement (SHALL):** Every period report cites the currently measured margin of error with reference to the Reading Error Matrix (docs/error-matrix.md): the discrepancy rate among records adjudicated in the independent second-reading round is **11.5%** (machine-to-machine agreement — the LOWER bound of the margin of error; the upper bound will be published with the human-labeled gold set). A badge cannot be presented as if this margin of error were zero (SHALL NOT); the momentum threshold (3.3.4) was chosen taking this margin into account. [Editor's note, September 29, 2026: 11.5% is the value measured on July 14, 2026; the current value is published in docs/error-matrix.md.]

**5.5 Prohibition of self-assessment (SHALL NOT):** Carbonbench itself, its affiliates, or legal entities associated with its operators cannot be badge candidates.

---

## 6. Sector annex framework

**6.1 Purpose.** For each subsector, an annex that is appended to the period report and generated AUTOMATICALLY is published. The annex places the badge results in the sector context; the single document + per-sector template-annex architecture (operator decision, July 2026) is made normative by this clause.

**6.2 Annex content template (SHALL — minimum in every annex):**
a) **Snapshot summary:** number of eligible companies n, total number of records, "competition opened/not opened" status (min n=8);
b) **Medians:** medians of the Accuracy Estimate, Transparency Score, and Target Quality + quartiles (p25/p75);
c) **Physical intensity bands (in sectors where applicable):** the `config/fiziksel_yogunluk.json` (fiz_yog_v1) bands with their source citations — e.g., crude steel: IEA Iron & Steel Technology Roadmap 2020 (global avg. ~1.4 tCO₂e/t; EAF ~0.3–0.7 · BF-BOF ~2.2) · cement: GCCA/IEA Cement Technology Roadmap (~0.6 tCO₂e/t) · primary aluminum: IAI statistics. In defining the intensity denominator (activity metric), the activity metric of the relevant **SASB industry standard** is taken as the basis (2.9); in heavy industry, a note on the scale bias of absolute/revenue-based percentiles is required;
d) **Period winners:** the results of the four badge types (with the rationale, if not awarded) + the Sector Leader sensitivity note (3.1.4);
e) **Universe notice:** "Universe = report publishers; silent polluters are out of scope" (Model Card §3).

**6.3 Production note (informative):** The annexes are produced by a **badge period generator** (planned module: `pipeline/rozet_donem.py`; 0 tokens, deterministic) that shares the data path of the existing sector league page generator (`pipeline/lig.py` — 17 sector static pages + JSON-LD): period snapshot → Clause 3 criteria → per-subsector annex HTML/MD + period report + snapshot summary seal. The embedded SVG badges (`pipeline/rozet.py`) gain a period-labeled variant for winners. Because the generator output is published together with the input snapshot summary, anyone can reproduce the annexes.

**6.4** Adding sector-specific fields to the annex template (e.g., a financed emissions note for finance, PCAF tiers) is done in a version window, and the same template version is applied to all sectors (SHALL — parity of criteria across sectors).

---

## 7. Limitations and disclaimer

**7.1** The scores that are the input to badges are **statistical plausibility estimates**; they are NOT an audit/assurance opinion, an ISO 14064-3 verification, a rating, or investment advice (Model Card §1, §6). A badge does not change this nature; it only documents periodic relative superiority.

**7.2** The data source is ONLY publicly available company disclosures and official public records (EPA GHGRP, EU ETS/EUTL, the official SBTi list, official grid EFs). Carbonbench does not perform on-site measurement; a consistently false disclosure can produce a consistently false score — and potentially an unearned badge (CB-P §2.3-a). This risk is limited by the cap mechanisms and official anchors but cannot be eliminated.

**7.3** "Sector Leader" and similar titles express the relative position within the subset of the relevant subsector that **publishes reports and is within Carbonbench's scope**; they are not a claim of absolute leadership across the sector as a whole. The universe is selection-biased, and this statement is visible on badge pages (SHALL).

**7.4** Badge images may be embedded on third-party sites; the embedding party shall use the badge together with its period label and profile link, without modification (SHALL). Presenting an expired period badge as if it were the "current status" is a violation of the terms of use.

**7.5** The statement "100 points is not a claim of completeness" (CB-P §9.9.1) also applies in the badge context: no badge certifies the completeness of a disclosure or the adequacy of the company's climate performance (including an SBTi alignment judgment).

**7.6** This English edition is the reference text of this Protocol. It replaces the Turkish text of July 20, 2026 in that role; if the two differ, this English text prevails.

---

## Appendix A — Badge criterion ↔ normative reference mapping table (normative)

This table maps, in an auditable manner, which external normative framework and which internal CB-P provision
each badge criterion rests on (in the pattern of CB-P Appendix A). There are no "decorative references": no
standard not cited here plays a binding role in the badge criteria.

| Criterion element | CB-B clause | External reference | Internal basis |
|---|---|---|---|
| Scope 1+2 magnitude and scope definitions | 3.0, 3.1, 3.4 | GHG Protocol Corporate Std. (2.1) | CB-P §6.5 canonical model |
| In-band disclosure requirement (location-based) | 3.1.2-b, 3.4.1-a | GHG Protocol Scope 2 Guidance (2.2) | CB-P §6.2.7, §9.5.1 |
| "Not a verification opinion" distinction | 7.1 | ISO 14064-3:2019 (2.3) | Model Card §1, §6 |
| Sector intensity comparison and annex medians | 3.1, 6.2-b/c | ESRS E1-6 (2.4) | CB-P §5.2.2, fiz_yog_v1 |
| Target Quality component | 3.1.1 | ESRS E1-4 (2.4) · SBTi (2.7) · TCFD legacy (2.10) | scoring.json target_quality_v2 |
| US coverage base rationale | 2.5–2.6 | SEC climate rule (with stay note) · CA SB 253/261 | — (coverage footnote) |
| TSRS/KGK universe framework | 2.8 | IFRS S2 (2.8) | tr_harvest/KGK archive pipeline |
| Annex metric selection + intensity denominator | 6.2-c, 6.4 | SASB industry standards / SICS (2.9) | CB-P §5.2.1 NACE bands |
| Disclosure-structure reading alignment | 2.11 | GRI 305 (2.11) | Extraction templates (GRI-code trap filtering) |
| Cross-consistency evidence (optional) | 2.12 | CDP public responses (2.12) | Evidence set (MAY); does not enter the score/badge directly |
| Momentum noise threshold (+5) | 3.3.4 | — | Reading Error Matrix (11.5% lower bound) |
| Formula freeze calendar | 4.4 | — | Model Card §5 (January/July) |
| Dispute grounds and record persistence | 5.3, 4.5 | — | CB-P §11.5, §8.6.1 |

## Appendix B — Terms (informative)

**period snapshot:** The set of active records sealed as of 23:59:59 UTC on the last calendar day of the period (4.3); the sole input to badge computation.

**eligible company:** A company that meets all of the Clause 3.0 prerequisites; the subsector n count is made only with eligible companies.

**subsector:** In v1.0, the SECTOR_MAP group of 17 (1.5); updated by version note upon the transfer to taxonomy v2.

**composite score (R):** The output of the Sector Leader formula (3.1.1); it is used in no other context and is not shown as a score on profiles — it is solely a periodic ranking tool.

**momentum (Δ):** The difference in the Accuracy Estimate between two consecutive period snapshots (3.3.1), with the formula effect netted out.

**period report:** The sealed publication containing the period results, the snapshot summary, the version imprint, the sensitivity table, and the margin-of-error statement (4.3, 5.4).

**Launch Edition:** The permanent label of the monthly-period badges specific to the launch year (4.2); not comparable with the quarterly series.

## Appendix C — Open items (v1.1 candidates, informative)

1. **Subsector taxonomy v2 transfer:** Upon the transition to the finer taxonomy with SICS↔NACE mapping, an empirical re-evaluation of the min n=8 threshold in fine slices (the risk that the competition never opens in small slices ↔ the outlier risk of lowering the threshold).
2. **Momentum composite:** The option of defining Most Improved on the composite R instead of on the Accuracy Estimate alone (so that Transparency improvement is also rewarded); in v1.0, a single metric was preferred for simplicity.
3. **Band-width normalization:** Using a direct upper bound on σ_ln,T (table published) instead of the Accuracy Estimate ≥ 70 proxy threshold in In-Band Accuracy.
4. **Empirical derivation of the Sector Leader weights:** Deriving the 0.5/0.3/0.2 vector from public anchors, in parallel with the CB-P §9.4.2 constant-derivation program; until then, the sensitivity table (3.1.4) is the mandatory compensating mechanism.
5. **Verified-claim badge variant:** A "claimed profile" mark in the badge SVG for claimed companies — solely as visual metadata, without affecting the independence principle (5.1).
6. **Multilingual annex:** Further language editions via the Phase 4 translation program; number/formula invariance is subject to the T10 language integrity test.
7. **SEC rule status monitoring:** Updating the stay note in 2.5 in every version window; reclassifying the reference if the rule becomes final or is vacated.

---

*— End of CB-B v1.0. Comments and objections shall be submitted, citing the clause number, through the Methodology Board Secretariat channel.*