Skip to main content
MG01MusicGOAT Index

Behind the list

Greatness is messy. Our method is not.

Music GOAT Index turns a famously subjective argument into something you can inspect, challenge and remix — without pretending taste is a fact.

MGI model 3.3.0July 2026 · MGI v3.3 archive6 factors counted2 context lenses

Artists

1,000

Ranked with one shared model

Core signals

6

Visible in every score

The ground rules

A ranking you can argue with.

The index is a point of view with inspectable evidence and visible judgment, not a divine verdict. These rules keep the argument honest.

01

One score is never the whole story

Every artist is judged across several dimensions, so a giant career cannot be reduced to one chart position or sales total.

02

The weights are visible

The all-time list has a published recipe. In GOAT Lab, you can move the weights and see exactly how the order changes.

03

Evidence has a confidence level

Strong public records, informed estimates and uncertain claims are treated differently — and labeled where they appear.

The scorecard

What actually counts?

The all-time model balances reach, impact and staying power. Sales and touring remain separate and currently unavailable until artist-specific evidence supports them — commercial scale is useful context, but not the same thing as greatness.

01Evidence-gated

Commercial

Commercial tier, sales evidence, certifications and scale.

Sales is unavailable until a reviewed artist-specific observation record supports a reproducible range and score.

0220% weight

Public attention

Long-horizon editorial cultural-footprint judgment. The frozen v3.3 compatibility score historically used a retired short-window Wikimedia Pageviews proxy; it remains provisional until replaced in a reviewed edition.

Public attention is a provisional long-horizon editorial judgment and makes no measured-observation claim; Current alone links the short pageview window.

0320% weight

Influence

Curated cultural and peer influence tier.

Influence is an editorial judgment, supported by notes about reach, legacy and impact on other artists.

04Evidence-gated

Live

Live/touring reputation and durable concert draw.

Live scale is unavailable until reviewed, attributable artist-specific observations support a reproducible range and score.

0515% weight

Longevity

Active years, persistence and legacy after the first peak.

Longevity is a provisional compatibility score with structured career dates where available; its normalization and manual review record are not yet published.

0615% weight

Catalog

Depth, spread and durability of the catalog.

Catalog strength is a provisional compatibility assessment supported by structured release history where available; normalization and review records remain incomplete.

0715% weight

Current

Short-window current-interest context derived from a dated Wikimedia Pageviews observation; it is not listening, sales or artistic quality.

Current presence uses recent public-attention context and should always be read with its date and source.

0815% weight

Peak

How dominant the strongest era felt relative to peers.

Peak is an editorial assessment of how completely an artist defined their strongest era, not a directly measured fact.

Evidence, without theatre

We show uncertainty instead of hiding it.

Public records first

Strong claims begin with sources people can inspect: charts, certifications, archives and established reference data.

Estimates stay labeled

Where no definitive number exists, the uncertainty remains visible instead of presenting a range as a measured total.

Checked before publication

Visitors browse a stable edition of the ranking. New research is checked before it can change a public score.

See the evidence policy

The short version

Scores are comparable. Taste is still yours.

Use the index to discover the strongest case, find where two legends differ, or build your own model. Disagreement is part of the product.

FAQ

Questions people ask about the ranking

Why do older artists not automatically lose because streaming did not exist?

The ranking uses normalized components rather than raw platform counts, so older and newer artists can be compared through era-aware signals.

How often is the site updated?

The public site uses versioned data snapshots with provisional evidence labels. Future updates are checked before publishing; daily automatic ranking changes are not enabled at launch.

Can artists request corrections?

Yes. The contact/correction flow accepts structured submissions and keeps review metadata private until a reviewer acts on it.

Does a normal page request call external music APIs?

No. Public pages load from versioned data snapshots. External source checks happen separately from visitor browsing.