EDITIONS vs VARIANTS, THE DISTINCTION EVERYONE ELSE BLURS
This is the one that quietly ruins most catalogues, and I will die on this hill. I have lives to spare. An edition is a tier of the same release: Standard, Deluxe, Exclusive, the convention-only colourway. A variant is a different version of the character: a new armour mark, a battle-damaged sculpt, a season-two costume. They are not the same axis. Mash them together and “Deluxe” ends up filed next to “Mark 7” as though they were the same kind of thing. They are not, and here they never are.
HOW WE READ TIME · ANNOUNCEMENTS, PROMISES & RELEASES
Time is the question I am asked most, dear resident, usually in the form “when will mine ship?”, and it is the question a catalogue is worst at answering, because “release date” is five different events wearing one coat. So the Archive keeps them apart, each at the precision it was actually known, and I will not have a year written as the first of January to make a calendar look fuller.
GLIMPSED is the month the Keep first sighted a piece in a maker's teasers. It is not an announcement, and some glimpses never come to anything. ANNOUNCED is the earliest defensible public reveal, as a day, a month or a year, and it is never promoted to a finer grain than the source gave. PRE-ORDERS OPENED is a separate, proven event; it is never assumed from an announcement. THE MAKER'S WINDOW is the first expected window a maker gave for a figure, captured once and never overwritten, so a delay cannot quietly rewrite what was first promised. CURRENT EXPECTATION is the window the maker stands by today, which may be a day, a month, a quarter or a span of months; it is a promise, not a shipment. CONFIRMED RELEASE is a historical fact: it needs dated evidence that the figure entered distribution (a retailer's arrival notice, the maker's shipping notice, an in-hand review, dated coverage, an archived page) or an Archivist's attestation, and the entry says which.
Two rules follow that collectors deserve in writing. A window that has closed with no confirmed release in the Archive reads WINDOW CLOSED, with the release date not yet confirmed in the Archive; and when collectors here already own the figure, it says so first, IN COLLECTORS' HANDS, because then it is our records that are behind, not the figure. It is not called late and it is not called unreleased, because the most common reason is that nobody has yet filed the evidence, and I will not accuse a maker on the strength of our own backlog. And a release is the moment a figure entered distribution, somewhere; it is not the day a particular shop ships your order. A figure can be released in one region and still show as a pre-order at your retailer for weeks. The Archive cannot see your retailer's warehouse and does not pretend to.
Time since announcement counts from the announcement to today, at the announcement's own precision: exact days when the announcement is a day, “about so many months” when it is a month, nothing at all when it is only a year, and nothing once the release is confirmed. It is a fact about the calendar, not a judgement: ten days after an announcement is not a delay.
The maker's historical line. Where a maker has enough history, the figure page carries one sentence of context: the typical time from announcement to confirmed release, across that maker's records. The rule is strict. A record counts only when both the announcement and the confirmed release are known to the day, the release is confirmed by evidence (never a window), it falls on or after the announcement and no later than the day of the reading. The interval is whole days. The headline is the median, because release times are skewed and an average is pulled about by the few very long waits; the middle half (the 25th to the 75th percentile, interpolated the way PostgreSQL's percentile_cont does it) says how wide the ordinary case is. Month-level records are a separate, coarser reading and are never pooled with the exact-date one. On a maker's page that broader reading counts records whose announcement and confirmed release are each known at least to the month, and compares calendar months: both dates are read as the first of their month, so a figure announced and released in the same month counts as zero, and the result is given in approximate months, never as exact days, because a release known only as “June” did not happen on the first of June. We say “announcement to release”, and never “pre-order time”, because the Archive does not yet hold enough pre-order opening dates to measure that interval honestly.
A maker earns the line only above a threshold: under twenty exact-date records nothing is said; from twenty to forty-nine the line is shown and called a limited sample; from fifty it is shown plainly. The sample size and the snapshot day are always printed beside the number. Note that the records are Archive entries, not necessarily independent release campaigns: sibling editions of one figure can share a chronology. And the line is history, never a forecast. A maker's median of 433 days tells you how that maker has behaved; it does not tell you when your figure will arrive, and the sentence says so every time.
The Archive is continuously researched and historical gaps are still being filled, so these readings improve over time. A later snapshot never rewrites what an earlier one claimed; it is simply a later reading, dated.
Kollectorium · 65, Rue de la Vallée, L-2661 Luxembourg City, Luxembourg