Methodology
Most anime ranking lists don't show their work. This one does. The rubric, the weights, the score scale, and what Anime Codex deliberately refuses to measure — all on one page.
The show-your-work thesis
There are a thousand “Top 10 Anime of All Time” lists on the internet. Almost none of them tell you what criteria they used, what weights they applied, or why one show beat another on a specific axis. You see a ranking, you don't know what to argue with, and the discourse circles forever.
Anime Codex starts from the opposite premise. Every show is scored against six explicit criteria with published weights, and every score carries a written critique explaining exactly what earned that number. You can disagree with a score. You can't say it didn't justify itself.
That's the contract. Everything below is the substance behind it.
The six criteria
Story & narrative
Plot structure, pacing across arcs, payoff vs. setup, whether the show earns its emotional turns or leans on shock value. Drift in the middle, sticking the landing, internal consistency under pressure.
Character writing & growth
Whether characters change in ways the story earns. Internal logic, motivation that holds up beyond convenience, supporting cast that exists beyond the protagonist's gravity well. Villains who reflect the show's themes vs villains who exist to be punched.
Themes & emotional resonance
What the show is actually about, underneath the surface premise. How well it makes you feel what it's trying to make you feel. Whether the themes hold up to re-reading or dissolve the moment you stop watching.
World-building & power-system uniqueness
The world's rules, how consistently they're applied, and whether anything about the setting or power system is genuinely new or just well-executed familiar. Shows that break their own rules to win fights take real damage here.
Animation & direction
Studio craft. Episode-level direction, key animation, fight choreography, the use of stillness, color, framing. The difference between “animated” and “directed.” Studios get praised when they earn it, named when they don't.
Cultural impact
How much the show moved the medium. Did it set a template other shows now copy? Did it land outside Japan in a way that changed what people expected anime to be? This is the lowest-weighted criterion deliberately — cultural footprint is real, but it's not story quality.
The weights are per genre
A 60-episode seinen psychological thriller and a 12-episode shoujo romance are not playing the same game. Evaluating both against identical weights would flatter shonen and unfairly tax the others. So the rubric is constant, but the weights shift by genre:
| Genre | Story | Character | Themes | World | Animation | Cultural |
|---|---|---|---|---|---|---|
| Shonen | 25 | 25 | 15 | 15 | 15 | 5 |
| Seinen | 30 | 25 | 20 | 10 | 10 | 5 |
| Shoujo | 20 | 30 | 25 | 5 | 15 | 5 |
| Josei | 20 | 30 | 25 | 5 | 15 | 5 |
| Kodomomuke | 15 | 20 | 15 | 15 | 25 | 10 |
- Shonen — roughly even split. Story and character carry, but world and animation matter because shonen lives on power systems and fight choreography.
- Seinen— story and themes weighted up. Most seinen worth ranking is doing something with the medium that lighter genres aren't, and the criteria reflect that.
- Shoujo— character and themes are the centerpiece. World-building is often deliberately stripped down because the show isn't about its setting.
- Josei— same weights as shoujo. Adult women's manga adaptations live and die on character interiority and emotional honesty; world-building rarely carries.
- Kodomomuke— animation and cultural footprint weighted up. Children's anime succeeds or fails on visual craft and how durably it lands across generations.
Cross-genre score comparison is therefore noisy. A 9.0 shonen and a 9.0 seinen aren't scored on identical weights. Use the per-genre ranking inside each category for the cleanest signal.
Disagree with the published weights? You can drag the six sliders yourself and see how the rankings shift. Local to your browser — doesn't change anyone else's view.
The score scale
Scores run 1.00 to 10.00. They aren't curved; they aren't calibrated to a normal distribution. They mean what they say:
- 9.5–10.0 — generational. The medium advances because this show exists. Almost no shows get here.
- 9.0–9.4 — exceptional across the rubric. The kind of show whose worst criterion is still good.
- 8.0–8.9 — excellent with specific visible flaws. Most fan-favorites land here once you actually look at every criterion.
- 7.0–7.9 — good with one or two criteria dragging. Often shows whose hype outpaces their execution on something specific.
- 6.0–6.9 — noteworthy for one criterion, deficient on another. Reading the per-criterion breakdown matters most at this tier.
- < 6.0 — doesn't earn its place in the conversation. Included anyway when the recognition threshold demands it.
What this deliberately doesn't measure
A score on Anime Codex is not influenced by:
- MAL member count. Popularity is a recognition threshold for inclusion, never a scoring input. Many shows in the catalogue ranked here are popular; their popularity isn't why they scored what they did.
- Reddit and Twitter sentiment. The discourse around a show changes monthly. The rubric doesn't.
- Nostalgia. Some shows that defined a generation score lower than younger shows that learned from them. That's the rubric working, not failing.
- Sales / box office. How well a show monetized has nothing to do with whether the writing earns its turns.
Inclusion rules
To enter the catalogue, a show must clear two bars:
- Magazine of origin. The source manga ran in a major magazine of the relevant genre (Weekly Shonen Jump, Bessatsu Shonen Magazine, Weekly Shonen Sunday, Monthly Shonen GanGan for shonen; Big Comic Spirits, Young Magazine, Afternoon, Young Jump for seinen; Bessatsu Margaret, Hana to Yume, Ribon for shoujo; Kiss, Chorus, Be Love, Monthly Flowers, Feel Young for josei; and so on).
- Recognition threshold. At least one of: 100K+ MAL members, a major streaming license, 50+ episodes or multiple seasons, or significant Western cultural footprint.
The cuts are deliberate. A show being good is not enough; it has to be widely-known enough that disagreement about its ranking is actually a conversation people are having.
One genre-specific exception: kodomomuke.
Kodomomuke (children's anime) is the one demographic where the anime IS the primary transmission medium, not the manga. Foundational kodomomuke franchises like Pokémon, Pretty Cure, Digimon, and Ojamajo Doremi are original anime works. Others — Crayon Shin-chan, Chibi Maruko-chan, Sazae-san — originated in non-kodomomuke magazines but are received by the public as kodomomuke. Strict manga-only would gut the catalogue. So for this one genre, the rule expands: the inclusion test is whether the anime is broadcast and received as kodomomuke, not whether the manga ran in a children's magazine. Recognition threshold still applies.
The AI in the loop
The per-criterion justifications were drafted using Anthropic's Claude model against a fixed prompt and a fixed rubric, then spot-checked for accuracy and consistency. The prompt version is stored with every ranking so any score can be traced to the exact generation that produced it.
What the AI is not doing: deciding the rubric, choosing the weights, picking which shows to include, or making editorial calls about what counts as a flaw vs. a feature. Those are editorial decisions. The AI generates the language; the framework and the calls are human.
The score that gives Demon Slayer Season 1 a 7.22 is not the AI being mean. It's the AI applying weights and definitions decided in advance, to a show whose animation carries writing that the rubric reads as mediocre. The score is a consequence, not a verdict.
Disagree with the rankings
This is a structured opinion, not an objective measurement. The rubric is one critic's framework applied consistently across a catalogue. Reasonable people will rank shows differently, and the project is more useful when those disagreements are specific — “the world-building score on this show is wrong because of X arc” is feedback that can shape future revisions. “Your taste is bad” is not.
The site is in active build mode. The catalogue now spans shonen, seinen, shoujo, josei, and kodomomuke. The criteria and weights are calibrated, not carved in stone. If a ranking feels wrong, a criterion needs rethinking, or a show belongs in the catalogue that isn't there yet, tell me.