Register · Check 0030 · External
Does the department's own calculator apply the correct standard deduction?
It does not, under the old regime. Found by computing the same return with a clean-room reference implementation.
Outcome
Divergent
Published
02 September 2026
It does not, under the old regime. Found by computing the same return with a clean-room reference implementation.
Held back deliberately, and not because it is unfinished
This check concerns a computation published by a government department, and the result is divergent. The material is complete — the inputs, both computations, and the arithmetic showing where they part. It is not published because a finding about somebody else’s system should be put to them before it is put to the internet, and that has not yet happened. Publishing first and notifying afterwards would make this an accusation rather than a check.
Check 0031 is the standard a complete entry is held to.
What is established today
- The divergence is real and reproducible: the same return, computed by a clean-room reference implementation, differs from the departmental calculator under the old regime.
- The cause is identified to a single statutory quantity — the standard deduction applied under section 16(ia).
- The reference implementation shares no code, no tables and no constants with the system it disagrees with. It was written from the statute.
What would move it to full publication
- 01The finding is put to the department in writing, with the inputs, both computations and the reasoning, and a reasonable period allowed for a response.
- 02Their response — including no response, which is itself a recorded outcome — is part of the entry.
- 03A named individual outside ProServiq re-runs the comparison and reaches the same result.
- 04Editorial review confirms the entry describes a computation and not a person.