---
title: "For lawyers: what the rating asserts, and what it refuses"
description: "A TrackForge rating is a building inspection, not a price index. It asserts what is registered, where and as of when, from named evidence, under a published rule book. It is not a title opinion, not a valuation, not calibrated to yield, and it discloses who did the fixing."
image: "https://trackforge.studio/docs/img/trackforge-social-card.png"
---

> Documentation Index
> Fetch the complete documentation index at: https://trackforge.studio/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# For lawyers: what the rating asserts, and what it refuses

This page is for the solicitor, counsel or diligence lead who is handed a TrackForge rating and has to decide what weight to give it.

## What the rating asserts

A TrackForge rating asserts, for each work the owner attests is theirs, and for each rights position the owner attests holding:

1. **What is registered, where, and as of when.** At each collecting route in scope, whether the owner is recorded with the right share and party, judged from a dated reading of a named source.
2. **Whether the work is unambiguously identified**, whether its ownership is documented and consistent with what the societies hold, whether its income is consistent with its usage where statements were supplied, and what structural risks are recorded against it.
3. **What could not be checked**, work by work, with the reason, and how much of the catalogue that leaves unassessed, by weight and by count.
4. **The rule book** the letter was computed under, by version and by hash, and every floor and coverage step that capped it.

It asserts these things deterministically: the same evidence rows under the same rule book give the same tuple. A party relying on the rating can have it recomputed under a separate agreement.

## What the rating refuses

The method is **a building inspection, not a price index**. It will not:

- **Give a title opinion.** The title component reports whether ownership is documented and whether the societies' records agree with the documents. It does not opine on whether the documents are valid, enforceable or complete. The owner's attestation proves the owner's claim; it fixes the population and decides no state.
- **Value the catalogue or calibrate to yield.** The letter says nothing about price, multiple or expected income. The method is not tuned so that letters predict returns, and the weights that rank works inside a catalogue rank them by usage, not by money, with income able only to raise a weight.
- **Compare catalogues of different shape.** A publishing catalogue and a recording catalogue are surveyed on different position classes; a UK heritage catalogue and a US streaming catalogue carry different routes. The letter is comparable within a rule-book version and a position class, with the routes and materiality disclosed, and no further.
- **Guarantee future collection.** A sound registration is a registration in order as of the reading's date. It is not a promise that the society will pay, or pay on time.
- **Score an opportunity, or let money move the letter.** Income not yet earned is listed, never scored. Measured amounts are disclosed and never change a score; value estimates sit in the Fix plan, outside the rating. The letter cannot be raised by prospects.
- **Grade what is not the owner's.** A work the owner is recorded on but whose income belongs to a third party under an agreement is removed by the attested grading scope, disclosed, and never scored.

## Independence and who did the fixing

TrackForge also fixes catalogues: it prepares and sends registrations, opens claims and chases societies. A rating of a catalogue TrackForge has worked on is therefore disclosed as such.

- Every rating event carries `remediated_by`: who did the fixing between this rating and the last. Its values are none, TrackForge, Ortpur (a company connected with TrackForge), the owner, or a third party.
- A before-and-after pair of ratings links only under the same rule-book version, so a remediation cannot be made to look better by a rule change.
- A comparison across rule-book versions carries a bridge note saying what changed.
- The rule book is versioned, committed by its hash before the first rating under it, applies no per-client customisation, and leaves no discretion in the scoring: the letter follows from the rows.

## What "not checked" means in a legal reading

A not-checked state is a recorded limit of the sources, never a judgement and never a choice. Treat it as absent evidence, neither for nor against. The rating's coverage fields say how much of the catalogue it covers; a letter that was withheld for coverage names the shortfall.

## Where the rules are

The rule book is one JSON artefact; its SHA-256 over canonical JSON is the methodology hash the tuple carries. It is published by that hash, timestamped on Bitcoin before the first rating under it; the [method statement](/docs/certification/rating/method-statement) sets out the same rules in words. The current version, v2.4, is a draft under calibration, and TrackForge issues no letter under it.

Source: https://trackforge.studio/docs/certification/rating/scope-and-independence/index.mdx