Home · Accuracy & corrections

Accuracy & corrections

If we publish a wrong number, we correct it visibly and record the change. This page is the public correction log for every TaxMetria jurisdiction and release.

Report a possible error

Found something that looks wrong? Tell us — we investigate every report. To speed up the investigation, include:

  • The URL of the page
  • The input you used (gross income, jurisdiction)
  • The figure you expected and where you got it (a primary source is best)
  • Your name and contact — we may need to follow up

Report address: corrections@taxmetria.com (placeholder until the production contact is set up)

Correction log

Every change to a public TaxMetria page or data record
Date Release Severity Summary
2026-09-16 tl-2026.1 n/a Initial release of US Federal 2026 + UK 2026/27 calculators. No outstanding defects at launch.

Known limitations

These are documented gaps in the current model. They are not bugs; they are intentional exclusions. Each one will become a tracked feature in a future release.

UK Personal Allowance taper (£100,000 – £125,140)

Applied correctly. The audit trail shows the effective Personal Allowance at any income level. Verified against HMRC guidance and pinned by boundary tests at £99,999, £100,000, £100,001, £125,139, £125,140, and £125,141.

US state income tax

No US state has its own calculator yet. The model is federal only. State-by-state calculators are on the public roadmap.

UK Scottish rates

Scottish residents have a different NSND band structure. The England/Wales/NI calculator must not be used for Scottish taxpayers. A separate Scottish calculator is in research.

OBBBA-era US senior deduction

The $6,000 additional standard deduction for taxpayers 65+ introduced by the One Big Beautiful Bill Act is not modeled in v1. Requires age input and additional testing.

How a correction is published

  1. Report received → triaged within 48 hours.
  2. If material, the data release is frozen and a patch release is prepared.
  3. The patch is added to the correction log here with a clear "before / after" entry.
  4. Affected pages are republished; visitors who shared the affected URL are informed where possible.
  5. The regression test suite gains a permanent test case pinning the corrected behavior.