How Hidden Temple Verifies a Checklist
The evidence rules behind every Hidden Temple checklist: printed numbers read off real items, sources parsed twice, and corrections published in full.
One rule, and everything else follows from it
A row on this site claims an identity only when a photograph of that physical item prints its own identity: a number band, a name plate, or a typed back. Not the National Pokédex order. Not the box art. Not the card either side of it in a spread. Not a sibling printing of the same set. Those four shortcuts are how almost every wrong checklist on the internet got wrong, because each of them is right often enough to feel safe. Where an item carries its own printed height, weight or category line, that gets cross-checked too, and that check has caught every misread found so far.
Unknown stays unknown
An unidentified row is published as unidentified. It says identity pending, evidence needed, and it stays visible so that a gap is legible rather than papered over. A set is never completed by inference from its own numbering scheme, and two rows are never merged because they look like the same Pokémon. The same species genuinely appears at two different numbers in several sets. Where a whole set is unfinished, some lines go further: a karuta box whose card list is not complete shows a locked tile with no list and no photo slots at all, so that a part-finished list cannot be mistaken for a finished one. That costs search traffic and it is still the policy.
Printed numbers get read twice, and at magnification
A printed number is read off the item, and then read again from a different photograph of a different physical copy. This is not caution for its own sake. In the typeface Amada used on its 2003 menko, a printed 5 and a printed 8 are not separable at ordinary magnification, and a row published from two board photographs read at normal size can be wrong while looking perfectly confident. So numbers are cropped and re-read at four and twelve times before they are bound, and a digit that is washed out by flash glare is recorded as unread rather than guessed from its neighbours.
Official sources get parsed twice as well
A printed total from a maker or a retailer is a claim until it has been read on two physically different copies of the thing that prints it. The count on the Game Boy Pocket card set is an example. The document that prints that total is the retail display band the maker made for it, and the figure on this site rests on two separate photographed specimens of that band, one lying flat beside a card grid and one square on and close enough that the numerals stand about forty five source pixels tall. Both print 4 box types, 151 card types and 5 body colours in their own three bullets. The same discipline applies to release totals: the menko figure on the Kakumen pages is taken from the reverse of two physically different sealed lottery boards from two different sellers, both printing Part.1 全100種類 above a table reading 70 Normal, 25 Rare and 5 Special. And where a number could have been filled in from sequence instead, it is not: a DP Retsuden sticker was published at 472 only because a dedicated list for that Part names 472 as Palkia between 471 Dialga and 473 Heatran, since the older source for it printed a question mark and sequence is not evidence.
Listings are not evidence. Photographs are
An identity is taken from the photograph and never from the title above it. Sixteen sealed menko boards were re-read in one pass, and one of them was titled as Plusle while its own photographs show number 081 printing Minun. If titles were trusted, that row would be wrong on this site today. The same rule applies to source spelling: where a collector checklist printed a species name with an apparent typographical slip, the corrected official name was published and the source's spelling was recorded alongside it as an alias rather than silently dropped.
Printings are never merged
Two boxes that share a name are not one product. The Diamond and Pearl karuta is four separate annual printings and this site keeps four separate lists for it, chosen one at a time, because a kana filled in from the wrong year is a fabrication dressed as a fact. The Advanced Generation karuta is two printings. Best Wishes is three. XY is two. Amada's first sticker set was printed twice and both runs are documented as runs. On the menko side the rule is sharper still: the two Kakumen parts interleave across numbers 001 to 201, so a number range never attributes a card to a part. Only a part-badged board or pack photographed in the same frame as the card can do that.
Every set says what its rows rest on
A set whose rows were verified carries an evidence note that says what was read, on what, and what that reading does not prove. It is written to be uncomfortable. The Kakumen Part 1 note states that a combined archive photograph confirms the shared numbering across both parts but does not show which of the two releases supplied each card, so unproven rows are not assigned to Part 1. The Game Boy Pocket note states that one number rests on a front read and a dex-ordered grid rather than on the printed band, because the band's own digits are glare washed at the resolution available. Disclosing the weak link is the point. A reader who knows which rows are load-bearing can check them, and a reader who cannot see that has to take the whole list on faith.
Corrections are published, not quietly swapped
On 2026-09-12 this site changed a published row, and the change is on the page it affects. Kakumen Retsuden Part 1 first published Azurill at number 084. The original reading came from two board photographs read at ordinary magnification, where a printed 5 and a printed 8 in that typeface are not separable. Three independent physical copies, read at higher magnification, all print 054 beside the same ルリリ name plate: the sleeved sample menko on two part-badged boards cropped and read at four times, and a loose card lying flat in a fifty card spread cropped and read at twelve times, where the open top of the middle digit finally separates a 5 from an 8. Two things then had to happen. The row moved to 054, because only the printed number was wrong and the species and the release did not move. And number 084 went back into the unresolved pool rather than being filled from the numbering scheme, recorded as unread rather than as vacant, because nothing photographed in that pass prints 084 at all. That second half is the part that costs something, and it is the part that makes the first half trustworthy.
The photographs belong to the people who took them
Almost every photograph on this site was taken by a collector who owns the item. Each one stays bound to its credit, and the contributors page names them. No third-party image is ever copied into this archive. When a marketplace listing or a maker's own product photograph is what proves a row, the address of that source is recorded in the research file and the picture stays where it is. That is why some rows here are verified and still show an empty photo slot: the proof exists, and permission to reprint it does not.
What this site deliberately does not publish
Prices and values. Not because they do not matter, but because a price is a snapshot and a checklist is a record, and mixing the two makes the record worse. Grading population reports and price trackers do that job properly and this archive points at them rather than competing with them. What is here instead is the printed number, the printed name in both languages, the finish, the maker, the year, and an honest statement of what each of those rests on.
See it working
Three places on this site where the standards above are visible rather than asserted. The Kakumen Retsuden Part 1 evidence trail, which carries the board readings, the printed release total and the 084 to 054 correction in full. The Game Boy Pocket card set basis, which names the two display bands behind its 151 total and says plainly which single row rests on a weaker read. And the Diamond and Pearl karuta page, where four annual printings are kept apart and the unfinished one is locked instead of filled.
Kakumen Retsuden Part 1 Game Boy Pocket card set Diamond and Pearl karuta
Questions about these standards
What should a collectible checklist include?
For a printed collectible: the number as printed on the item, the name in the original language and in English, the finish or variant, the maker, the year, and a statement of what each row rests on. A checklist that gives a number and a name but never says where either came from cannot be checked, and a list that cannot be checked is a rumour with a table around it.
How do you look up a Japanese Pokémon card or sticker?
Start with what the item prints on itself: the number, the Japanese name, the maker mark and any code on the packaging. Those identify a product. Artwork, colour and the Pokémon alone do not, because makers reused all three across different sets and years. Every set page here is indexed by printed number and carries the Japanese name alongside the English one for exactly that reason.
How do you identify Pokémon menko?
By format first (square kaku menko, round maru menko, or patchin), then by the maker mark and the printed number. For the Kakumen line, the part cannot be read from the number, because the two parts interleave across 001 to 201. Only a part-badged board or pack photographed with the card can attribute a part, which is why this archive will leave a menko unattributed rather than assign it by its number.
What happens when Hidden Temple gets something wrong?
The correction is published on the page it affects, with the evidence that overturned it and the date. The Azurill row on Kakumen Retsuden Part 1 was published at 084 and corrected to 054 on 2026-09-12, and the number it vacated went back into the unresolved pool rather than being filled from the numbering scheme. Corrections here are additions to the record, not edits to it.
Why are some checklists here incomplete or locked?
Because the alternative is guessing. An unidentified row says so, and a karuta box whose list is not finished shows a locked tile with no list at all rather than a part-filled one. Both cost this archive traffic and both are deliberate. A gap you can see is information; a gap filled by inference is damage.
Can I submit a photo, and who owns it?
Yes, through the upload slots on any set page that is open. You keep your photograph. It stays bound to your credit, your name goes on the contributors page, and nothing is published until an admin has reviewed it. No third-party image is ever copied into this archive.