About

The convention belongs beside the answer.

Quantilenote checks one chosen percentile of a small entered sample using an explicit Type 7 linear-interpolation rule.

Sample-percentile conventions can differ. An unexplained result can hide whether a tool selected an observation, interpolated two values or treated endpoints differently. Here the named convention, exact rank, bracket order positions and weight remain visible alongside the answer.

A focused exact model

Duplicates count as separate observations. Decimal parsing, ranking and interpolation use BigInt/rational arithmetic; the result includes a reduced fraction and complete terminating decimal, even when output needs more fractional places than the input.

The guide distinguishes a sample calculation from population inference and explains why ties do not guarantee that the chosen fraction of observations is strictly below the result. It also states the accepted grammar and finite size bounds rather than guessing how to clean missing values.

Local, free and inspectable

No registration, upload or conversation is required. Optional text/JSON notes include every input value and the convention. This release has no analytics, advertising or consent-platform loader; hosting and optional email remain separate.

Independent exact probability-knot interpolation, parser/boundary fixtures and shipped-handler tests support the source release. Actual browser and public hosting evidence are separate steps. These checks establish behaviour under the stated model, not population accuracy or decision suitability. Read examples, privacy or use contact with invented test data.