Artists, albums, songs, and genres — all scored, no popularity contest.
Turn a criterion off, or make it count for more or less. Changes apply immediately and are remembered on this device.
Untick a genre to hide every artist in it from the table below, the leaderboard, and search — the rest of the page works as usual. Not remembered between visits.
Sort or filter by any criterion using the controls above, or tap a column heading to sort by it directly. Tap How this is scored above the table for the full rules, and Adjust weights to change how much each criterion counts — right after you do, any row that moved shows how many places, up or down. Artists in the overall top 100 have an i next to their name: tap it for a breakdown of how their score is built, including why it isn't higher on any criterion scored below the top of the scale. Check the box in the Compare column for up to 3 artists to see them on a radar chart. Click a sorted column again to go back to the overall ranking. A ◆ next to a name marks a hidden gem — Great tier or higher on merit alone, but a Fame & sales score of 2.5 or below; check Hidden gems only above to browse just those.
Each genre's top artist under your current weights, and how the genre scores on average. Tap a row to jump to that genre in the table below.
| # | Genre | Top artist | Top /100 | Genre avg /100 | Artists |
|---|
This is the most subjective panel on the site. It scores the 26 genres themselves, not the artists in them, on eight genre-level criteria — a completely separate, equally-weighted exercise from the artist scoring above, with its own numbers. It doesn't affect, and isn't affected by, anything else on this page. See "Best genres" in How this is scored for the full disclosure.
Core genres and the broader catch-all buckets (International Pop, International Rock, World & Chanson) are scored the same way, but a bucket built for taxonomy convenience rather than a single musical identity naturally scores lower on Innovation and Catalogue depth. That's a property of the bucket, not a verdict on the music inside it.
The 12 core genres only, one dominant line of descent per genre — real genre history has far more cross-pollination than any tree can show; this picks a single defensible parent for each so the shape stays readable.
Turn a criterion off, or make it count for more or less. Changes apply immediately and are remembered on this device. Separate from the artist and genre weights above.
Also the most subjective panel on the site. It scores 158 landmark albums, not artists or genres, on eight album-specific criteria — a completely separate exercise with its own numbers, not derived from anything else on this page. Every one of the scored artist table's top 100 has at least one album here. It's still a curated selection, not an exhaustive ranking of every great album ever made, so an absent favourite outside that top 100 is a gap, not a verdict. Weights work the same way as the artist table's: turn a criterion off, or count it for more or less. See "Albums" in How this is scored for the full disclosure.
Instrumental or near-instrumental albums (Kind of Blue, A Love Supreme) hold a neutral 3 on Lyrical depth rather than being penalised for having nothing to score there.
Turn a criterion off, or make it count for more or less. Changes apply immediately and are remembered on this device. Separate from the artist, genre, and album weights elsewhere on this page.
Also the most subjective panel on the site. It scores 103 individual songs, not albums or artists, on eight song-specific criteria — a completely separate exercise with its own numbers, not derived from anything else on this page. Every one of the site's 26 genres has at least one song here. It's a curated selection, not an exhaustive ranking of every great song ever recorded, so an absent favourite is a gap, not a verdict. Weights work the same way as the artist and album tables': turn a criterion off, or count it for more or less. See "Songs" in How this is scored for the full disclosure.
Not part of the scoring above. No criteria, no weights, no score, no effect on any artist's tier or ranking — a separate, factual list of the highest-grossing tours on record, ranked by gross. Figures are not adjusted for inflation, so older tours are understated relative to recent ones — sourced from Billboard Boxscore, Pollstar, and Guinness World Records reporting, current as of when this was last updated. Limited to tours with a reliably documented gross AND a matching ticket-count figure; some very high-grossing tours are left out because no verified ticket count could be sourced for them. Click a column to sort by it.
Not part of the scoring above. No criteria, no weights, no score, no effect on any artist's tier or ranking — a curated, chronological list of videos picked for how they pushed the format itself.
Not part of the scoring above. No criteria, no weights, no score, no effect on any artist's tier or ranking — a curated, chronological list of cover art picked for becoming as recognisable as the music inside it.
Not part of the scoring above. No criteria, no weights, no score, no effect on any artist's tier or ranking — just curated, chronological side lists for things the ten criteria don't cover on their own. Nothing here is exhaustive, and nothing here is ranked against anything else.
This describes the full method behind every score on this page: what each criterion measures, how the total is built, what's covered in detail versus estimated, and how the list and its rules came to exist. It states rules and history, not conclusions about whether the result is fair. It's written so it can be copied and given to an independent reviewer, human or AI, with no other context, to judge the scoring on the actual method rather than on a summary of it.
Lyrics how well the words do their job, judged in the language they were written in, never in translation. Simplicity used on purpose, for provocation, humour, minimalism or a chant, is credited when it works. Generic cliché is not.
Composition melody and songwriting by the artist or band. Artists who wrote almost none of their own material are capped at 2.5, however good the songs they sing.
Musicianship instrumental and production skill, judged on the recordings as a whole, including backing bands and arrangers.
Vocals emotion and character of the voice, or flow for rappers, not technique. 5 is unforgettable and deeply expressive, 4 strong and distinctive, 3 effective, 2 functional, below 2 minimal or purely textural.
Innovation how new the sound was. This rewards being first, not being widely known, so sorting by it can surface a genre pioneer few people outside it have heard of, ranked above famous names who came later.
Influence effect on other artists, not sales. Gives extra weight to reach achieved without singing in English, on the stated reasoning that such reach comes without the advantage of the language most of the global market already runs on.
Live stage performance. 5 is legendary shows, 4 strong and reliable, 3 average or inconsistent, 2 weak or rarely performed. For artists active mainly before 1950, this counts at half weight, since reliable accounts are scarce, unless they had a long, well documented live career.
Vision how distinct and fully realised the artist's aesthetic world is, visual, conceptual, or sonic, independent of whether it was the first of its kind. Reviewed in detail only for the 200 highest-scoring artists at the time it was added. Everyone else holds a neutral 3 until reviewed.
Catalogue combined strength and staying power of the whole catalogue, not just the hits. Consistency and longevity are merged into one score, so a great short catalogue and a great long one can score the same. Capped at 4 for artists who chose to release three albums or fewer, and at 4.5 for four to six. A career cut short by death or a split is not capped.
Fame & sales chart success, record sales, and mainstream name recognition, nothing else. Weighted the same as every other criterion by default (1×); can be turned down or off in the panel above if you'd rather scores stay independent of popularity.
Each criterion is scored 0 to 5 in steps of 0.5. The weight shown in the panel below is the actual weight used in the formula — there is no hidden base weight sitting underneath it. At the default 1× across the board, all ten criteria count exactly equally; nothing is quietly weighted more until a viewer moves a slider themselves. Wanting Innovation and Influence to count for more than the rest — because they measure lasting artistic impact rather than reception — is a defensible view, and it's available as the Critics' picks preset below, but it's an opt-in choice, not the unlabelled default.
The panel above lets a viewer multiply any criterion by 0, 0.25, 0.5, 0.75, 1, 1.25, 1.5, 1.75, 2, 2.5, or 3 (0 = off). The total is the sum of (score × current weight) across all ten criteria, with Live's weight halved for a pre-1950 artist not on a fixed list of 15 names with well-documented live careers. The maximum possible total (MAXT) is 5 × the sum of the current weights. The displayed score out of 100 is total ÷ MAXT × 100, rounded to two decimal places. A viewer's weight choices are stored on their own device and used for no one else. Because every criterion starts on the same 1× footing and scores move in half-point steps, many artists land on the same total. Those ties are broken criterion by criterion (see "Tie-breaking and sort order" below).
Tier cutoffs, as a fraction of MAXT: Legendary 86%+, Elite 82%+, Great 77%+, Notable below that. Because MAXT changes with the weights in use, the cutoffs are recomputed the same way whenever weights change. These were loosened from an earlier, stricter set (88/85/81) after review found the top bands too narrow to fit artists who clearly belonged there, and found "Notable" catching a wide range of artists that included some well above merely notable.
The Hidden gems only checkbox above the table flags artists whose quality outran their recognition. It recomputes a separate merit score from only the nine non-Fame criteria (using your current weights on those nine, but always excluding Fame & sales regardless of what you've set that weight to), and flags an artist as a gem when that merit score clears the Great-tier bar (77%+) while their actual Fame & sales score sits at 2.5 or below. A gem is marked with a ◆ next to its name whether or not the checkbox is on, so it's visible while browsing normally too. Artists on a quick estimate are never flagged, since a gem claim needs the confidence of a full review behind it. Six of the dataset's genres (French Chanson, Arabic & Middle Eastern, South Asian, K-pop, International Pop, and Fado & Portuguese) currently return zero gems under this definition: every artist curated into those smaller, more selective genres already scores above 2.5 on Fame & sales within their own category, which is a fact about who made the cut for those genres, not a flaw in the gems logic.
1,014 of the 1,087 artists in the dataset have a full, individually reasoned score on every criterion, including Fame & sales. The remaining 73 have a faster, less individually scrutinised estimate, marked with a ~ next to the name; for these, the table shows the tier but not an exact number, and Fame & sales defaults to a neutral 3 until reviewed. The detailed group started as the 500 highest-scoring artists and was later extended to cover every artist who mainly sings in English, regardless of score, once it became clear that review depth was tracking language more than merit. The 73 left on a quicker estimate are, with very few exceptions, artists who don't mainly sing in English.
Every ranking on this page (artists, albums, songs, and genres) uses the same cascade. First, the weighted total, recalculated every time a weight changes. If two totals are exactly equal, the criteria are compared one at a time in a fixed order, and the first one that differs decides. Each is counted at its current weight, so a criterion turned off also drops out of tie-breaking. If every criterion matches, the two share a rank number and are listed alphabetically. Sorting by a single criterion puts that criterion first, then follows the same cascade.
The order: for artists, Innovation, Influence, Composition, Vocals, Musicianship, Lyrics, Catalogue, Vision, Live, then Fame & sales. For albums and songs, Innovation, Influence, Songwriting, then Cohesion (albums) or Vocals & performance (songs), Production & sound, Lyrical depth, Emotional impact, then Longevity. For genres, the criteria in the order they appear in the table. Lasting impact comes first, the craft of the work next, and reception last.
The Best genres panel scores the 26 genres themselves, not the artists in them, on eight criteria: Innovation, Influence, Musicianship & craft, Lyrical & emotional depth, Catalogue depth, Longevity, Live culture, and Cultural reach. Each is scored 0 to 5, equally weighted, summed to a total out of 40, and shown as a score out of 100 — the same 0-100 scale as the artist table above, but a completely separate exercise with its own numbers, not derived from any artist's score in any way. It doesn't affect, and isn't affected by, anything else on this page, and there's no weight adjustment for it yet.
This is the most subjective section on the site. Scoring one artist against another is already a judgment call; scoring an entire genre — decades of different artists, scenes, and eras compressed into eight numbers — is a bigger one, made by one person with AI help, the same way the rest of this page was built (see "How the list and its rules were produced" above). It's included as an honest attempt at something genuinely hard to do fairly, not because the numbers are more defensible than an argument at a bar. Disagree with it; that's the expected outcome.
The Albums panel scores 158 curated landmark albums on eight album-specific criteria: Songwriting, Cohesion, Production & sound, Innovation, Influence, Lyrical depth, Emotional impact, and Longevity. These are deliberately different from both the ten artist criteria and the eight genre criteria above, because a single 40-70 minute work asks different questions than a career or a genre does — Cohesion (does it hold together as one listen, or play like a track dump) and Production & sound (the specific studio craft of this one record) don't have artist- or genre-level equivalents on this page. Scored 0 to 5, summed to a total out of 40 (or less if a criterion is weighted down), shown as a score out of 100. Same separation as the other two: its own numbers, no effect on anything else here.
Unlike Best genres, this panel does have its own weight adjustment: the Weights button above the table works exactly like Adjust weights on the artist table, letting a criterion be turned off or multiplied up to 3×. It's a fully separate weight system from the artist table's — Innovation here only affects album totals, not artist totals — stored under its own key on the same device, defaulting to every criterion counting equally (1×) until changed.
A hundred and fifty-eight albums is a broad, curated selection chosen for range across genre and era, not a claim that these are definitively the only ones that qualify — there is no album equivalent of the artist database's goal of comprehensive coverage, and some artists appear more than once where more than one of their albums is widely considered a landmark in its own right (Radiohead, Bob Dylan, Kendrick Lamar, Michael Jackson, and others). Every one of the site's 26 genres has at least one album on this list, including the regional and international genres that got zero representation in an earlier version of this panel — most of those still only have one, since the biggest gap in this canon is genuine breadth outside the Anglo-American mainstream, not a lack of worthy candidates. Earlier versions of this panel also missed a large share of the scored artist table's top 100 outright — the count went from 100 to 158 specifically to close that gap, and every artist in that top 100 now has at least one album here (a few, like Louis Armstrong and James Brown, sat as high as #2 and #3 overall with no album representation at all). High artist rank still doesn't guarantee an obvious album pick — the two exercises are independent and album selection is its own judgment call — but a top-100 artist with zero presence here was a real gap, not a defensible curatorial choice. Each album carries a genre tag (the same taxonomy as the Genre filter on the Artists table) so the panel can be filtered by it. An instrumental or near-instrumental album (Kind of Blue, A Love Supreme, Giant Steps, Bitches Brew, Selected Ambient Works 85-92, and others) holds a neutral 3 on Lyrical depth rather than being scored down for something it was never trying to do, the same convention this page already uses for Vision scores awaiting review (see Coverage above). Built the same way as Best genres: one person, AI help, openly disclosed as a judgment call rather than a settled fact.
The Songs panel scores 103 curated individual songs on eight song-specific criteria: Songwriting, Vocals & performance, Production & sound, Innovation, Influence, Lyrical depth, Emotional impact, and Longevity. These sit closer to the album criteria than the artist ones, since a single song and a single album are both one piece of work, but Cohesion doesn't apply to a three-minute track the way it does to a 40-minute record, so it's replaced here with Vocals & performance — how the specific vocal or lead performance on this one recording holds up, independent of the songwriting underneath it. Scored 0 to 5, summed to a total out of 40 (or less if a criterion is weighted down), shown as a score out of 100. Same separation as the other three rankings on this page: its own numbers, no effect on anything else here.
Like Albums, this panel has its own weight adjustment: the Weights button above the table works the same way, letting a criterion be turned off or multiplied up to 3×, stored under its own key on this device, independent of the artist, genre, and album weights.
A hundred and three songs is a curated selection chosen for range across genre and era, not an exhaustive ranking of every great recording ever made — a few artists appear more than once where more than one of their songs is widely considered a landmark in its own right (Bob Dylan, Metallica, Bob Marley and the Wailers, Édith Piaf, Amália Rodrigues, and others). Every one of the site's 26 genres has at least one song on this list. A handful of entries here — novelty and viral hits like Macarena, Dragostea Din Tei, and Aserejé — deliberately score low on Songwriting and Lyrical depth while still scoring well on Influence or Longevity, because cultural reach and lasting craft are different things and this panel is built to let scores disagree with each other rather than smoothing that over. Each song carries a genre tag on the same taxonomy used elsewhere on this page, so the panel can be filtered by it. Built the same way as Best genres and Albums: one person, AI help, openly disclosed as a judgment call rather than a settled fact.
The Language field records the language an artist mainly sings in, not their nationality or country of origin. An artist who performs mainly in a language other than the one associated with their home country is filed under the language actually sung (for example, a French band performing mainly in English is filed under English). An artist without one dominant language is filed as Multilingual.
The dataset covers popular music: rock, pop, hip hop, jazz, blues, soul, country, electronic, and adjacent genres, where one artist or band is typically the writer, performer, and voice at once. In classical music these roles are usually separate people, often centuries apart (a composer, a conductor, an orchestra, a soloist), so most of the ten criteria as defined would need to be scored against a different person each, or not at all. Classical music, and the period and body of work it covers, is also large enough that including it would require its own separate set of criteria and its own separate list rather than a subset of entries added to this one.
This list and its scoring rules were built through an extended, iterative conversation between one person and an AI assistant (Claude), not compiled from a fixed source or a single blind pass. Three points from that process matter for judging the result: some artists were added because that person asked about them by name rather than through an independent, systematic search for omissions; some individual scores were revised after that person questioned a ranking, without new biographical or discographical information being introduced as the reason; and coverage depth (see below) was at an earlier point unevenly distributed by region before being corrected. None of this is disclosed to argue the result is fair or unfair — it's given so a reviewer can weigh the method itself.
This needs a direct disclosure rather than a bibliography: no specific article, book, or database was looked up and cited for any individual artist's score. The scores come from an AI language model's general knowledge, built during training from a very large mix of text it cannot list or verify after the fact, and applied through the criteria and rules described above. There is no per-artist citation trail to publish, because none was kept, and none could be reconstructed after the fact with any honesty.
What can be said honestly is the kind of material that knowledge is drawn from in general, and each kind carries its own reliability and its own bias:
Professional music criticism (outlets such as Rolling Stone, Pitchfork, AllMusic, NME, The Guardian, and equivalent publications, absorbed in aggregate rather than as individual citations). Wide-ranging and generally careful about canon and reputation, but the field itself is disproportionately English-language and Western, so its judgement is more available, and probably more accurate, for Anglophone and European artists than for others.
Chart and sales figures (Billboard, IFPI, certification bodies). Objective and verifiable in principle, used loosely to shape the Fame & sales criterion. Reliable where formal tracking exists, weaker for markets and eras with less formal chart infrastructure, which again skews away from some regions and periods.
Reference and biographical material (encyclopedic sources, discography databases). Broad coverage and generally reliable for facts like release dates and discography size. Not a source of critical judgement, and coverage depth still varies by how much has been written about an artist in English.
Fan and critical consensus over time (best-of lists, retrospectives, aggregated listener sentiment). Useful for gauging how an artist's reputation has settled, but this is a contested, evolving canon, not a fixed truth, and it carries the same English-language, Western-leaning skew as the rest.
In short: the underlying knowledge leans toward being more complete and more reliable for English-language, Western, and already well-covered artists, and thinner for everyone else. That is exactly the pattern the Confidence score below is built to make visible.
Every artist now has a Confidence rating of High, Medium, or Low, shown as its own column. It measures how much individual review and how much reliable documentation likely sits behind a score, not how good the artist is: a Low-confidence artist could still be scored correctly, and a High-confidence one could still be wrong.
It's built from two things already on this page: whether the artist got the full, individually reasoned review described in Coverage above (rather than the faster estimate), and whether they mainly sing in English, used here as a rough, honest stand-in for how much English-language documentation is likely to exist to check the score against.
High: reviewed in detail and mainly sings in English. Medium: one of those two is true but not both. Low: neither, meaning a quicker estimate for an artist whose main documentation, if any, is less likely to be in English. This isn't a judgement on non-English artists or their work; it's a disclosure that they were checked against less material, for the reasons given under Sources above.
After extending detailed review to every English-language artist (see Coverage above), the breakdown is 919 High, 95 Medium, and 73 Low out of 1,087. The 95 remaining Medium-confidence artists are almost all detailed reviews of non-English artists: they got the same individual scrutiny as everyone else, but the English-language documentation used to check that scrutiny is thinner for them, and that's a fact about available material, not something more review time fixes. The 73 Low-confidence artists are non-English and still on a quicker estimate; unlike the Medium group, those can be moved up by giving them the same detailed pass, on request.
To be direct about the part of this that's uncomfortable: an artist who mainly sings in a language other than English cannot reach High confidence here, no matter how well-known or important they are within their own scene, and no amount of extra review time changes that on its own. That's worth stating plainly rather than leaving implied.
The reason is not that non-English artists are less deserving, harder to judge fairly, or less significant. It's narrower than that: the underlying knowledge this scoring draws on (see Sources above) is itself unevenly distributed, skewed toward English-language music journalism, English-language biographical writing, and English-language critical consensus. For an artist whose main coverage exists in Portuguese, Arabic, Wolof, or Turkish rather than English, there is less English-language material available to check a score against, whatever the actual quality of that artist's work or the depth of coverage in their own language. Confidence reports that asymmetry in available material. It does not report, and is not meant to report, artistic merit.
Two things follow from that. First, Confidence is a separate column from the score and tier: it never adds to or subtracts from an artist's total, so a Low-confidence artist's ranking is exactly as high or low as their nine scored criteria put them, untouched by this column. Second, this is the same asymmetry already disclosed elsewhere on this page (see "Built through a conversation with one person" and Sources above) restated at the level of an individual artist rather than the dataset as a whole. A reviewer auditing this page for bias should treat the Confidence column as a map of where the underlying material is thin, not as a second opinion on whether any specific artist belongs where they are.