Comparison

Hume's test case is a pairing nobody now remembers was ever a contest: John Ogilby, a workmanlike seventeenth-century versifier, against John Milton. His point was uncomfortable: a person acquainted with no better poetry might genuinely admire Ogilby — and the admiration would be sincere, felt, and wrong in a way the admirer has no way to detect. By comparison alone, he argued, do we learn to assign degrees of praise; whoever has seen only one kind of beauty cannot rank any.

Read that carefully and it is not a claim about how much music you have heard. It is a claim about what breadth gives you — degrees. The judge who has weighed many works can say that one is a little better than another and a third is far worse; the judge who has not is left with liking and not-liking, which is not a scale but a floor.

This page used to promise something else, and the correction is worth stating rather than hiding. It described an optional import of your streaming history and said that breadth was a fact about your listening rather than a skill anyone could test. That version needed a catalogue we would have had to license and a taxonomy we would have had to invent, and it measured what you had been exposed to rather than what you could do with it. The version that shipped measures the thing Hume actually named.

It reuses a test you have already taken. The Prestige Test asks you to rate sixteen clips blind on a scale of eleven whole numbers, then rate them again with names attached. Those ratings are already on your device, so comparison costs no new clip and no new tap. Two things come out of them: how many of the eleven degrees you actually landed on, and how many pairs you ordered one way blind and the opposite way once the names were time — counting only pairs where the labels pushed both clips the same direction, so a prestige label cannot be the explanation.

Neither number is a mark out of anything. The count is read against what an indifferent rater would produce rather than against the top of the scale, because rating sixteen clips at random already lands on about nine distinct values — the ceiling is reachable by accident, and a reader measuring themselves against it is measuring themselves against nothing.

What the professionals do with their own scales

Assigning degrees is not an eighteenth-century abstraction; it is the daily work of music criticism, and its central embarrassment is how few degrees anyone uses. These are quoted as a reference point and never as a target — nobody here is scored against a critic, for a reason given below.

Pitchfork: The scale runs from 0.0 to 10.0 in tenths, which is a hundred and one places a record can land. Across more than 18,000 reviews published between January 1999 and January 2017, the mean score was 7.0. Most of those scores lie between 6.4 and 7.8. Scores ending in .0 appear nearly twice as often as scores ending in .1 — the reviewers avoid the decimals their own scale offers them.

Robert Christgau's Consumer Guide: The Consumer Guide's letter grades ran from A+ down to E−. From 1990 he used fewer letter grades for records below B+, replacing the bottom of his own ladder with honourable mentions and the categories Choice Cuts, Neither and Duds.

Ours is 0 to 10, whole numbers only — eleven places to put a clip. The instrument asks only how many of them you used — not whether you used the right ones, because on this question there is no right one.

Why it never scores you against a critic

The obvious version of this instrument compares your ranking with a famous reviewer's and tells you how close you got. It is not built, and it is not going to be. The Prestige Test exists to measure how far a prestigious name moves your judgment. A product that measures how far a famous name moves your ratings cannot also give you credit for agreeing with a famous critic. So critics here set the spread and never the answer, and the instrument never says a reader is wrong.

The honest limit, last, because it matters more than anything above it: these clips were never spaced out by quality. They were chosen for licence clarity and for genre spread, so nobody knows how far apart they truly sit. If they really are close together, hearing them that way is the correct answer — and this instrument cannot tell that apart from a listener who hears everything as much the same. It reports what you did with the scale. It does not grade your ear.

The measurements themselves, with their formulas and their caveats, are published in the Lab.