The Completion Desk exists because almost nobody writes about the second half of a long game. This page explains how we count, what our terms mean, and the rules we keep so that one report can be compared against another.
Everything is written down as it happens, on paper, in a ruled book with one page per session. Each entry records the in-game date, the real elapsed time, the objectives completed, and a one-line note about whether anything was new. It is a deliberately low-technology approach and we have tried the alternatives.
Recording as you go matters more than the medium. Reconstructing a hundred-hour playthrough from memory afterwards produces a narrative — usually a flattering one — rather than a record. The ledger disagrees with our memories often enough to justify itself several times per game.
A task type is defined by the operation the player performs, not by the fiction wrapped around it. Two quests that both require you to travel to a marker and defeat everything there are the same type even if one is a bandit camp and the other is a haunted mill. Conversely, two superficially similar cases are different types if the mental operation differs.
This is the single most contested definition on the site and we apply it consistently rather than conveniently. Where a classification is genuinely arguable, the report says so and shows both counts.
Whatever the game itself describes as full completion: 100% on its own counter, every achievement it lists internally, or every entry on its own checklist. Where no such state exists, we define one and publish the definition.
A unique player operation. Reskins, difficulty variants and reordered steps do not create a new type. A change in the decision the player must make does.
The last hour at which the game introduced something new. Recorded from the ledger rather than from feeling, and often earlier than any of us would have guessed.
Our recommendation for where a reader should deliberately finish, chosen to land on a satisfying resolution rather than simply at the point of decline.
Three people who between them have finished more long games than is defensible. We do not publish names, biographies or contact routes: none of it improves a report, and a completion desk that invites correspondence would spend its time on correspondence rather than on games.
What we do disclose is the part that affects a report. Every piece states who played it, how many hours they logged, whether they had played anything else by the same developer, and whether they enjoy the genre. That last one matters more than most sites admit.
Counting errors happen and we fix them in place, with the change noted on the page and the original figure left visible. A rating changes only where a correction is large enough to alter the conclusion, which has happened twice and is recorded on the rating page.
Games with no end state and no plausible stopping rule, and live-service games whose requirements change while we play. Neither can be completed in any meaningful sense, and pretending otherwise would make our figures worthless.
