Power Scaling: A Reader's Guide

The three primary inputs: feats, hype, and statements

Feats are what a character is shown to do. Saitama splits a planet in a dream sequence; Gojo holds a domain unsupported. Feats carry the most weight because they're the least ambiguous — they're on the page or on the screen, they're calibrated against an environment, and they require less inference. The serious caveat is that feats are written for narrative, not for power-scaling. A character can have done something massive once and never repeat it.

Hype is what other characters say a character can do. The villagers say Madara is unmatched; Whitebeard says Roger and Garp would have killed each other if they fought seriously. Hype is useful because it sets the audience's expectations, but it's softer than feats: characters in fiction are often wrong, and hype can be propaganda, fear, or politeness.

Author or in-universe statements — databooks, magazine interviews, side material — sit between hype and feats. They're often used to settle arguments, but they can be retconned, mistranslated, or framed conditionally. Power-scaling debaters usually treat databooks as evidence, not proof.

Scaling chains: how arguments get built

Most power-scaling claims aren't direct: they go through a chain. A typical chain looks like this. Character A defeated Character B. Character B beat Character C off-screen. Character C fought Character D evenly. Therefore Character A scales above D.

The chain only holds if every link does. Three links break in predictable ways:

• Plot-driven outcomes. A fight ended a certain way because the story needed it to. The winner isn't always the stronger character; sometimes the writer needed someone to die to push the next arc. Jujutsu Kaisen, Death Note, and Game of Thrones all have famous outcomes that scaling chains then have to reconcile.
• Stakes mismatch. A character at full power versus a character holding back is not the same fight. Off-screen fights, in particular, are unreliable because the audience didn't see whether the loser fought seriously.
• Power growth. The character at the start of an arc is not the character at the end of it. Scaling chains that ignore the gap between "end of saga 4 Goku" and "mid saga 6 Goku" collapse on contact. Dragon Ball debates run on this footnote.

The ladder problem

Long-running shounen face a structural problem: the protagonist has to keep getting stronger to keep the story interesting, but every gain inflates the ladder. By the late saga, the early-saga ceiling characters are barely a footnote. Dragon Ball, Bleach, and One Piece have all had to do this; My Hero Academia manages it more conservatively.

Two consequences for power-scaling debates. First, "is X stronger than Y" is incomplete without a timestamp; comparing peak X to debut Y is a category error. Second, late-saga power inflation creates discontinuities that early scaling chains can't bridge — the character who was the ceiling in arc 1 might not even threaten the floor in arc 7. Honest debates note when a chain is broken by an inflation event and stop trying to extend it across.

Cross-franchise comparisons (Goku vs. Saitama, etc.)

Cross-franchise debates are the most popular and the least conclusive, because the two settings don't share a calibration. Saitama in One Punch Man works as a joke about how absurd shounen power inflation is — he is, by authorial design, beyond his setting's ladder. Comparing him to a character whose own setting doesn't include that joke is comparing two narrative devices rather than two characters.

Useful framing for cross-franchise arguments:

• Pick a fight rule. Are abilities "on"? Is prep allowed? Are the characters in-character or playing to win? Most disagreements are about the rules, not about the characters.
• Decide whose physics applies. Light Yagami in Death Note's setting can win against almost anyone; in another setting where his rules don't bind, he loses to a footsoldier. Where the fight happens often decides the answer.
• Acknowledge the comedic ceilings. Joke characters — Saitama, Bill Cipher in Gravity Falls, certain Toriyama-tier-mascot characters — aren't really fighting on the same axis. Cross-comparing them with serious shounen leads draws out the ladder, not the result.

Worked example: a clean and a messy debate

Clean: Itachi vs. Sasuke (Naruto, the actual canon fight). Both characters' feats are on screen. The fight resolved in canon, with the source material specifying the conditions and the outcome. The scaling chain is short: A vs. B, on-screen, full effort. Disagreements here are about reading interpretation, not about the underlying evidence.

Messy: Madara vs. all-out Itachi. Itachi has a higher feat ceiling than what he displayed in canon, because plot constraints (his illness, his goals) suppressed his real capacity. Madara has high feats and very high hype, but most of his fights had narrative restrictions of their own (sealing, time limits). The chain has to estimate what each character would have done at full output, which adds inference at every step. Two careful debaters can read the same evidence and disagree honestly.

The general lesson: a fight that happened on the page is easier to debate than one that didn't, and a single-character debate (was X holding back?) is easier than a head-to-head that never occurred.

Common mistakes

• Quoting a feat without its context. "He destroyed a planet" means little if the planet was a dreamscape, the destruction was metaphorical, or the feat was non-canon. Always check the medium and the canonical status.
• Conflating durability and power. A character who can survive being hit by X is not necessarily a character who can dish out X. The hit-taker and the hit-giver are two different ratings; databases like the tier-list community separate them.
• Treating PIS / CIS as a trump card. "Plot-induced stupidity" (PIS) and "character-induced stupidity" (CIS) are real phenomena, but invoking them every time a feat is inconvenient hollows out the debate. Use sparingly, and only with a clear narrative reason.
• Using one databook line as a verdict. Databooks are evidence. They aren't the source material itself, and authors can and do contradict them in later canon.
• Skipping "in character" vs. "optimal play". Light Yagami fighting in-character is a different opponent from Light Yagami fighting to win at any cost. Most of the famous "upset" arguments come down to this distinction.
• Comparing across mediums without translation. Game-mechanics damage numbers don't map cleanly to story feats; manga panels don't read at the same scale as anime adaptations. Comparing them as if they did inflates one side artificially.

How to read a power-scaling argument productively

A few habits make the genre more enjoyable and less tribal.

• Ask for the rule set first. If two people disagree about whether a fight is in-character or optimised, the argument can't progress. Pin down the rules before you pin down the answer.
• Separate "wins more often" from "wins always". Many character match-ups have a probability spread, not a guaranteed outcome. Treating the question as binary forces overconfident claims.
• Allow timestamps. "Adult Naruto in Boruto vs. mid-Shippuden Sasuke" is a real, answerable question; "Naruto vs. Sasuke" without a timestamp is several questions stacked on top of each other.
• Treat losses as data. A character losing to a story constraint isn't proof of weakness, but pretending the loss didn't happen isn't honest either. Note what the loss tells you and what it can't.

Where to go from here

For the rankings and tier lists themselves, see the tier lists hub — that page links out to specific franchise rankings. For the storytelling side of how rivals get written, see iconic rivalries in long-running series. For how archetype labels (mentor, foil, anti-hero) interact with strength readings, see the character archetypes guide.