Public evidence

math.percentage

Machine-checked verification for calculation version 1.1.2 · tier engine_tested

Open calculator API Capability Download JSON CVP Manifest MCP Review policy

CVP status

All seven modes; sampled numeric scales, signed inputs, exact zeros, denominator/domain errors, representable subnormals; nonzero conversion and output under/overflow rejected, including delta and average. Independent CPython Fraction reference; private280-case corpus.

Declared partition coverage: PASS · 9/9 declared partitions (of, is-what, of-what, change, adjust-up, adjust-down, difference, invalid-domain, xcal)
Versions: Calculation 1.1.2 · CVP protocol 1.0.0-proposed · Evidence 2026-10-21.percentage-complete-contract-v2
Verification revision: 2026-10-21.percentage-complete-contract-v2 · 2/2 property · digest 9cbd3aac385f

Verification

Legacy regression suite: 19/19 calculation tests · Production surface contract 6/6

Three distinct suites: educational examples on the calculator page, this legacy regression count (19/19 golden + boundary), and the CVP verification suite below. They are not the same number. Property, metamorphic, round-trip, O3, and interface statuses/checks are recorded in the CVP Manifest . SKIP (0/0) means that suite has no published vectors.

STALE · Build schema 1.1.2 matches page · production binding STALE · Semantic contract ✓ · Last attestation PASS · current evidence changed · re-attestation required · CVP STALE · attestation STALE — Production CURRENT withheld · Public/cache ✓ · Origin ✓

CVP verification suite

Protocol checks from the Evidence Manifest. Distinct from the legacy regression suite above. Verification revision 2026-10-21.percentage-complete-contract-v2. Calculation version 1.1.2. CVP 1.0.0-proposed. 253/253 golden · 7/7 CVP boundary · 34/34 invalid · 2/2 property · 2/2 metamorphic · 2/2 round-trip · 14/14 cross-interface · 3/3 cross-calculator · 6/6 CVP contract. digest 9cbd3aac385f.

Golden (oracle-backed) 253 / 253
CVP boundary 7 / 7
Invalid input 34 / 34
Property 2 / 2
Metamorphic 2 / 2
Round-trip 2 / 2
Cross-interface 14 / 14
Cross-calculator 3 / 3
CVP contract 6 / 6
Interfaces PASS · UI (SSR) / REST / MCP
Interface claim ui-ssr is query-result HTML, not a live browser session. Error-path engine·REST·MCP 7/7 (status, code, calculation_version). SSR compared on URL-canonical requested calculations; empty query is idle (not an error) and JSON-typed object/array inputs are REST/MCP-only.
Assurance Core

release artifacts aligned PASS
schema output required declared PASS
schema valid input executes PASS
engine success output validates schema PASS
schema rejects incomplete input PASS
rest success payload validates schema PASS

Fingerprint sha256:accc82aa83125aba50801a7e42b167e787b4d92eec083499f069575015858029

Input schema sha256:1c4027a3599acb763985ac7ecfd8ed4e411f0b0905aaff49598666c96c82f871 · Output schema sha256:62eabb18f2d5ccc6a7b0b8b561429cd32642cd156ab07623b3ecf2dea3f50b64 · Build-time SHA-256 digests prove published artifacts agree with each other (release_integrity.scope=workspace). Production attestation GETs canonical URLs twice (public/cached view and origin), including Capability HTML, via npm run attest:production --write (published at /.well-known/calculatorx-production-attestation.json) and fails if the public Capability page advertises a different calculation_version. This is artifact consistency, not an independent immutable release log.

Build & scope

Engine identity and declared limitations.

Tool: percentage
Engine: percentage
Engine build: 2026-10-21
Evidence core: 2026-10-21T10:39:28.506Z
Rendered: 2026-10-22T04:36:23.795Z
Last reviewed: —

Assumptions

Interactive evaluation runs locally in the browser. Shared URLs and REST evaluate the same versioned engine (query SSR).
Denominators (whole, old, average of A and B, P) must be nonzero where required.
Percent change uses |old| in the denominator (signed increase/decrease). Typically used with a nonzero baseline; old = 0 is undefined.
Percentage difference is defined for A ≥ 0, B ≥ 0, and A+B ≠ 0.

Limitations

Exact formulas of normalized binary64 inputs, output rounded once ties-to-even. Decimal inputs are rounded binary64, not exact decimal arithmetic.
Reject nonzero values that round to zero and overflow (including published delta/average); rescale on NUMERIC_RANGE.
Change uses |old|; difference needs nonnegative a,b and nonzero sum. Near zero baseline is ill-conditioned; shortest round-trip display is not significant-figure tracking.
At most512 characters per scalar; active fields only and aliases must agree.

Sources

Standards and references supporting this tool specification. Methods are listed separately — they are algorithms, not bibliographic sources.

BIPM — The International System of Units (SI), 9th edition — Units accepted for use with the SI — percent (symbol %), 1% = 0.01
Supports: Defines percent as the number 0.01 (one part in one hundred). The core relation P% of X = X × P/100 follows from that definition.
NIST Special Publication 811 — Guide for the Use of the International System of Units (SI) — Writing values of quantities — the percent
Supports: NIST usage of percent as 0.01 with the SI. Does not prescribe percent-change vs percent-difference conventions.
CalculatorX percentage conventions — Six phrase modes, |old| percent change, non-negative percentage difference
Supports: Percent change Δ% = 100×(new−old)/|old| (undefined at old = 0). Percentage difference Δ% = 100×|A−B|/((A+B)/2) on A ≥ 0, B ≥ 0, A+B ≠ 0. Increase/decrease is X×(1±P/100).
NIST — SI Unit Rules and Style Conventions — Percent and fractions
Supports: The symbol % represents the number 0.01. Thus p% = p/100 and a part of p% of base b equals b×p/100; inverse percentage equations are algebraic rearrangements. CalculatorX’s negative-baseline signed-change convention is separately specified, not attributed to NIST. Mathematical reference only; not implementation validation or an endorsement of CalculatorX.

Verification disclosure

Aggregate outcomes, scope and artifact commitments are public. Complete test inputs, expected values, execution records and oracle infrastructure are retained privately. This summary does not independently reproduce the full verification run.