This is the full text of the standard, as published. The version under change control is held at the public repository, which records every change and who made it. This page and that record are the same document. Read with jurisdiction profile EW-1.

The Calmly Resolve Assessment Methodology Standard

Published 22 August 2026 · v2.0, 3 September 2026 · jurisdiction profile EW-1 (England and Wales) · Tough Minds, Tender Hearts Ltd · Calmly Resolve

This standard defines how Calmly Resolve assesses individual consumer disputes. It is published so that anyone — a party, a funder, a court, an auditor — can check an assessment against it. Conformance under this standard means the methodology was followed; it never warrants the outcome of an individual assessment. Independence is carried by publication, simultaneous disclosure and an independent check of the published record (§4), not by an endorsing institution.

The standard is read with a jurisdiction profile. The standard says how an assessment is made. The profile (§7) supplies the country-specific inputs the assessment reads. Neither document is complete on its own. Every report and every line of the calibration record states both versions, written as v2.0 / EW-1.

Each section below is marked global or country-scoped. A global section applies unchanged in every country. A country-scoped section reads the profile in force and produces a different result under a different profile.

1 · Scope — global

The methodology applies to individual one-off consumer disputes where:

the amount in dispute is £200–£25,000
the counterparty is identified and reachable
the dispute concerns one transaction, not a pattern of conduct or a group claim

Disputes outside this scope are not assessed under this standard.

Monetary bounds are stated in the currency of the profile in force. The bounds above are the EW-1 bounds, in pounds sterling. Section 7 states how another profile restates them.

2 · Admission — the evidence gate — country-scoped

A case is admitted for assessment only when determinative documentary evidence is present at intake. The test is issue-bearing, not document-counting:

Identify the determinative issue — the question the case turns on (what was agreed; whether it was done; what it was worth).
Work back from that issue to the evidence held at intake. A document counts only if it bears on the determinative issue. Invoices, terms and payment records that do not speak to what the case turns on do not satisfy the gate.
Qualifying evidence classes: a signed contract or written terms; a payment record; dated photographs or an independent report of the defect; a written admission; proof of non-delivery; correspondence evidencing the disputed promise.
Value floor: cases below £500 are not admitted (assessment and resolution costs exceed any realistic recovery at that level).

The gate is evidential, not predictive. No model scores the likely winner at intake; admission asks only whether the documents can settle the question.

What this section reads from the profile. The value floor and the list of qualifying evidence classes are country-scoped values. The figures and the list above are the EW-1 values. A profile for another country states its own, and the reader takes them from that profile, not from this text.

Profile precondition (rule 2 of §7). No claim is admitted unless a published profile exists for the country in which the claim would be issued. A claim with no profile is refused at intake and the refusal reason is recorded as "no profile".

Coding discipline. Evidence presence is coded twice, independently, by different routes: a document-led pass (inventory of what is held) and an issue-led pass (work back from the determinative issue). Disagreement between passes resolves to the issue-led coding, with a written reconciliation note per case. Inter-coder agreement is measured (Cohen's kappa) and a codebook revision is required below κ 0.7.

3 · Assessment — the expert valuation (EV) report — country-scoped

Every admitted case receives an EV report with these sections, in order:

The dispute — parties, transaction, amount, outcome sought.
Evidence considered — a numbered manifest, documents only, stating what was not available.
What was agreed — findings from the contract, listing or messages, cited by evidence number.
What happened — findings of fact, cited; where the evidence is silent, the report says so.
Liability — the basis in the applicable rules, the reference class used, and the conclusion, with the determinative evidence named. The applicable rules are the consumer-rights baseline named in entry 7 of the profile in force. Under EW-1 that is the Consumer Rights Act 2015, read with the contract terms.
Quantum — a central figure and a range, with the basis stated (price paid, cost to cure, part-performance).
Settlement recommendation — a figure and a window, with one line on why it lands inside both parties' realistic outcomes.
Methodology and conformance statement — the standard version and the profile version applied, written as v2.0 / EW-1, and the conformance checklist (§6) completed.

Reference classes. Liability and quantum judgements are anchored to a reference class — dispute type crossed with value band — carrying documented base rates from resolved cases. Reference-class tables are maintained centrally and updated only from coded outcomes (§5).

Reference classes are per profile (rule 1 of §7). Every reference class belongs to exactly one profile and carries that profile's name. No reference class may draw base rates from more than one profile. Where a country has too few resolved cases to populate a class, the class is empty and the report says so. It is never filled from another country's cases.

4 · Independence — publication, disclosure and the checked record — global

Three mechanisms carry the independence of an assessment:

Published standard. This document is public and versioned. Every report states the version it was produced under, so any reader can test the report against the standard without relying on the assessor.
Both-parties disclosure. The full report is disclosed to both parties simultaneously. Neither side receives a different document. The report is the shared basis for settlement.
Checked record. The calibration record (§5) — predicted against realised outcomes by reference class — is published, and is checked periodically by an independent accountant under agreed-upon procedures (ISRS 4400) covering completeness of coding and the arithmetic of the published variance figures. The check covers the record, never an individual valuation.

No institution countersigns individual reports. Version 1.0 of this standard provided for institutional endorsement of each report; that provision is withdrawn.

5 · Calibration — the outcome loop — country-scoped

Every closed case is coded against its intake state within five working days: evidence-gate result, EV prediction, outcome, settlement value, time to resolution, counterparty response class, variance notes. Coded outcomes are the only permitted source for updating reference-class base rates. Predictions are compared to outcomes on a rolling basis; systematic variance in any reference class triggers a methodology review.

One profile per line, one profile per class. Every line of the calibration record states the standard version and the profile version it was produced under. The outcome loop updates the reference classes of that profile only. A coded outcome from one country never updates a class held under another profile.

Publication. The calibration record is published by profile. A published record states the profile it covers, so a reader can see which country's outcomes it measures without asking.

6 · Conformance — global

An assessment conforms to this standard when all eight of these are true:

the case was in scope (§1)
a published profile existed for the country of issue (§2)
the gate was applied as specified, with the coding discipline (§2)
the report contains all eight sections, with citations (§3)
the standard version and the profile version were both stated (§3, §4)
simultaneous disclosure occurred (§4)
every reference class used drew on one profile only (§3)
the outcome was coded against that same profile (§5)

Anyone may audit any assessment against this checklist.

7 · Jurisdiction profile — global

A jurisdiction profile is a named, versioned, published file. It holds the country-specific inputs that sections 2, 3 and 5 read. It is not prose and it is not an annex to this document: it is a separate file with its own version, published alongside this standard.

Naming. A profile name is a country code and a serial number, for example EW-1 for England and Wales. The standard version and the profile version move independently. A change to either is a change to the pair, written v2.0 / EW-1.

Required entries. A profile is complete when it holds all eight:

Path tree — the outcome states a claim can reach, and the transition rate into each one. Each rate states its source and whether it is researched or judgement-set.
Court and procedural fee schedule — the fee at each step, by value band, with an effective date and the published source it is taken from.
Undefended-judgment rate — the share of claims that reach judgment without the defendant contesting the merits, with the denominator stated.
Enforcement route — the steps available after judgment, in order, with the precondition and the cost of each.
Limitation period by claim class — the period, the date it runs from, and the statute that sets it, for each class of claim the profile admits.
Solvency registers — the register consulted for each legal form of defendant, named, with the conditions on which admission is refused.
Consumer-rights baseline — the statute or rules liability is assessed against under §3 section 5.
Card-rights baseline — the rights a buyer already holds before the cover exists, and the inputs that place a transaction in a residual-rights band.

Provenance marking. Every number in a profile is marked researched or judgement-set. A researched number names its published source and the date it was read. A judgement-set number says so, gives the band it is pre-registered within, and names what will measure it. A judgement-set number is never published as if it were researched.

Two rules.

Rule 1 — one profile per reference class. No reference class may draw base rates from more than one profile.
Rule 2 — no claim without a profile. No claim may be admitted unless a published profile exists for the country in which the claim would be issued.

Currency. A profile states its currency. The monetary bounds in §1 and the value floor in §2 are read in that currency, at the figures the profile states.

Change control. A profile change is versioned and dated like a change to this standard. Reports already issued keep the profile version they were produced under. The calibration record is not restated across a profile change; it carries the version each line was coded under.

Version 2.0, 3 September 2026 (v1.1, 2 September 2026; v1.0 first published 22 August 2026). Changes to this standard are versioned; assessments state the version and the profile they were prepared under.

Version note — what changed at v2.0

Nothing in v1.1 changed in substance. Every rule of v1.1 is still here and still says the same thing. The same scope, the same evidence gate and coding discipline, the same eight report sections, the same independence position, the same outcome loop, the same conformance checklist. No rule was relaxed, tightened or withdrawn.

What changed is that the country is now stated. Version 1.1 was written for England and Wales and did not say so. Its reference classes were built from resolved English cases, its liability clause named an English statute, and its money floor was in pounds. A reader in another country could follow every rule in it and still produce a wrong number. The numbers the rules read were English, and nothing on the page said so.

Version 2.0 makes the country a named setting the standard reads instead of an assumption it carries. Three sections are marked country-scoped (§2, §3, §5), three stay global (§1, §4, §6), and a new §7 defines the profile that the country-scoped sections read. The first profile is EW-1, England and Wales, published with this version. The two rules in §7 are new: they close the two ways a mixed-country number could otherwise be produced without anyone noticing.

The independence position of §4 is unchanged. There is no institutional endorser. Independence is the published standard, simultaneous disclosure to both parties, and an independent check of the published record.