Beast of Reincarnation Charms: One Rolled Mod, Not Six

The five lines are not on the same scale, and the grid hides it

Every wiki publishes charms as a five-by-three grid with a percentage in each cell, which invites the one comparison the data cannot support. Max HP at +70% and Attack at +7% are percentages of different bases, and nobody has published either base — so a reader cannot tell from the grid which is worth a slot on Koo, and neither can this site. What does survive is the comparison inside a line: what a III is worth over a I is a ratio, and a ratio has no units to disagree about. HP and Defense multiply by three and a half across their tiers. Skill Power multiplies by two. Upgrading a Skill Charm is worth, in its own terms, less than half what upgrading an HP Charm is — and no record can say that, because a record only knows about itself.

HP and Defense are the same three numbers

+20%, +40%, +70% appears twice: once on the HP line and once on Defense, identical at every tier. Two lines carrying byte-identical values is either a real design decision or one table copied into two rows by whoever transcribed it, and nothing we hold settles which. It is flagged rather than smoothed over because the difference matters to a reader: if it is a transcription artefact then one of those two columns is wrong, and the site publishing it does not say where its numbers came from.

Your charm has one rolled mod. The wiki lists the pool

This is the finding, and it comes from the only Charm panel anyone has read off the game itself. The panel prints “Boost Skill Power +20%” — exactly the published figure, which is the reason to trust the rest of the grid — and beside it a single extra effect, “Crit Boost 3”, where the wiki lists six for that row. So the Charm Effects column is the pool a charm rolls from, not the list a charm carries. Every guide reprinting that column as an item’s effects is describing something that does not exist, and a reader comparing their own charm against it will conclude their game is wrong. The same frame settles a spelling: the game writes Perfect Dodge, so “Perfecrt Dodge” is that wiki’s typo rather than a name.

They drop from enemies, and the farm exists after all

This section has been wrong twice in the same direction and both corrections are kept, because the shape of the error is the useful part. On 9 August it said no source gave an acquisition location for any of the fifteen. On 10 August Fervor’s lightning build guide supplied the mechanism — Charms drop from enemies in the later zones — and added, unprompted, that he had found no deterministic farm and would publish one if he did. On 22 August kamigame published one. It names a twin-blade golem near the camp 「山間の村前」 in 巨壁前棚田区域 that drops charms at a high rate, with a flower field beside it to assassinate from, and recommends the 隠密草生成 skill to make that easier. So the map exists; it was written in Japanese and no English source has it. What Fervor said about the *shape* of the problem still stands and the farm does not change it: he was hunting an ATK Charm III carrying a specific rolled mod, which is two independent rolls, and a route that produces unidentified charms quickly does not make either roll less random. It makes rerolling cheap — which is what the next section is for.

The rolled mod has a published pool now, with numbers

Two Japanese wikis publish the pool that single mod is drawn from, and AppMedia attaches a figure to every entry — the only source in any language that does. This is corroboration of this page’s central claim from two directions at once. Neither wiki was looking at the client panel that produced it, and both describe a charm the same way this page does: one primary line, plus one secondary effect. Fextralife’s six-per-row column is the pool, and here is what is in it. The two lists are not the same length — AppMedia names nineteen entries, kamigame thirteen, because kamigame collapses the six elemental variants into one row — so a mod missing from one list is not evidence it does not exist.

• The disagreement worth reading is on the primary line, not the pool. AppMedia ranks HP ★★ and argues it is nearly redundant once healing skills come online mid-run; kamigame ranks it ★★★★. Both rank Critical Rate first and Skill Power second. Nothing published settles HP, and this page does not average two tier lists into a third.
• AppMedia gives the one argument here that is a mechanism rather than a ranking, and it is the most useful sentence either wiki prints: Critical Rate is the only one of the five primary lines that cannot also be raised by attribute points or by cooking. Every other line has a second source; that one does not, so the charm slot is the only place it comes from. That is checkable against the cooking and attribute systems rather than being someone’s tier opinion.
• Its two crit figures reconcile, which is a small point in its favour: a Critical Rate charm plus cooking is quoted at 50%, and 80% is quoted when the rolled mod is also 会心 — and 会心 is published at +30%.
• Every number in this table is one wiki’s claim, read 2026-08-22, with no method or provenance behind it. None has been reproduced here, and the base-value problem in the first section is untouched by any of it.

Rerolling is a save-reload, and it turns on when the charm is identified

kamigame publishes the procedure and it depends on a piece of game behaviour worth knowing on its own: an unidentified charm is identified automatically the moment you touch a camp. So the reroll is: save at the camp *before* picking the unidentified charm up, take it, touch the camp to identify, and if the roll is wrong, quit to title and reload rather than saving. Repeat. Paired with the twin-blade golem route in the previous section, that is a loop for the secondary roll specifically. Two further constraints kamigame states and no English source does: an identified charm cannot be sold or discarded — it stays in the inventory permanently — and the current version has no favourite-marking or sort-by-effect function, so a long reroll session leaves you scrolling a list by eye. Neither claim has been reproduced here, and the sell/discard one is the kind of thing a patch changes quietly.