Primary
Current eCFR text
The governing regulation is the authority. Secondary material does not override it.
Trust Center
How RatingScope checks VA rating guidance before publication
RatingScope starts with the current regulation, checks it through an independent source channel, records ambiguity instead of guessing, and implements criteria as deterministic rules rather than model-generated percentages.
What you can verify
Evidence instead of trust claims
These are the controls behind the published guidance and the limits on what they prove.
Static
Deterministic criteria rules
Percentages come from hand-authored rule paths, not model-generated numbers.
Visible
Sources, assumptions, and gaps
Guidance identifies the rule area and keeps unresolved facts or ambiguities visible.
Reviewable
Implementation and source checks
Condition builds require a line-by-line source cross-check. Automated tests are not treated as regulatory proof.
Our verification principles
Controls in the verification standard
These controls define what research and implementation must establish before supported guidance is published.
Primary-source first
RatingScope researches the current text of 38 CFR through eCFR and sources its rating logic from published regulatory criteria. Secondary summaries are not used as authority.
Independently cross-checked
A second source channel corroborates the research. When a VA DBQ exists, it can confirm what is measured, but the regulation controls if the two do not align.
Official text stays identifiable
Official regulatory language, RatingScope's plain-English explanation, and any disclosed gap are presented as different things. A summary is never presented as the law itself.
Independent
RatingScope is not affiliated with, endorsed by, or part of the Department of Veterans Affairs. Optional donations do not unlock features or change a result.
Reviewable, not self-certifying
Condition implementation requires a line-by-line comparison with the source research. Automated tests can catch drift, but RatingScope does not treat a passing test as proof that a regulatory interpretation is correct.
The verification process
From regulation to published guidance
The research standard separates source authority, corroboration, implementation, and review so one check cannot stand in for all four.
Step 1
Retrieve the regulation text
Directly from eCFR, the federal government's current text at the time of review, not a cached summary.
Step 2
Cross-check against the VA's own forms
Where a VA Disability Benefits Questionnaire (DBQ) exists, we check that our description of what gets measured lines up with it. If a DBQ implies something the regulation itself doesn't say, the regulation wins.
Step 3
Flag any genuine ambiguity
Cross-references, undefined thresholds, or open-to-interpretation language are recorded as gaps that must not be silently resolved by software.
Step 4
Encode and cross-check the rule path
The condition logic is built as deterministic criteria and checked line by line against the source research. Tests support the check; they do not replace regulatory review.
Step 5
Schedule review and respond to discrepancies
The research standard calls for review at least annually and an immediate review when a source changes or a discrepancy is reported.
What we will never do
The hard boundaries
These aren't aspirations. They're limits we've built the product around.
- We will never generate or predict a rating with AI, or let a model guess at a percentage.
- We will never predict your claim's outcome or promise a specific approval, benefit, or decision.
- We will never give you legal or medical advice, or act as a replacement for an accredited representative.
- We will never guess on a genuinely ambiguous regulation. It gets flagged for a human decision instead.
- We will never ask you to upload your medical records to use this site.
I built RatingScope because I went through the VA claims process myself, and I know exactly how confusing "cite 38 CFR §4.whatever" is when you're just trying to figure out what your own paperwork means. This page exists to show how we check our work, including the boundaries of what the process can establish, instead of asking you to trust a label. If you find a discrepancy, please report it so we can compare it with the current source.