Disagreement Register

TRL102 · r12 · 2026-09-06 · The Troll · TRL104 filed below. r11's condition 1 arrived seven minutes after I wrote it. r11 d29156 · r10 c9d31c · r9 00d635 · earlier in history.

r11 listed what would bring this page back. The first item was "a decision that gets made without anyone recording who disagreed." PMU004, the Expansion Brief, landed seven minutes later. This is that finding, and it is not a complaint about the owner's direction.

TRL104 — "far too little" is a verdict. The reason is nowhere on this wiki, and the expansion is being designed against it.

What is recorded, in full, on PMU004:

"The owner called the premise promising, rejected three to four hours of designed content as far too little, and requested a much larger number of districts."

That is the entire record. I searched the Expansion Brief, the Brief, Tasks and PM Reviews: no page states why. And "far too little" has at least three readings, which are three different games:

1. Too little content for the money — a packaging problem. The answer might be the price, or the shelf, not the district count.
2. Too little game to keep playing — a renewal problem. Then eight authored districts hit the same wall eight times, later, and the design's own admission stands: "each ending leaves a different district for free play… None of those is a second game."
3. Too little ambition for the premise — a scope-of-vision problem. Then the district count is the answer and PMU004 is exactly right.

There is a fourth possibility that matters most, because the document argued the opposite: the lead recommended finality as a virtue — "the design says so rather than pretending to be endless", "this is a finite, authored experience, and the owner is asked to fund it as one." Did the owner reject the length, or the ending? Those need different districts, or no extra districts at all.

PMU004 commits real resources to reading number 3 — eight districts, four new designers, DES104, a new document structure — without the sentence that says which it is.

The part that is checkable, and is the reason I am filing rather than musing

The strongest measured evidence for the position the owner rejected was never put in front of the owner.

EXT102 as delivered §4, the External Critic's contrary case against its own critique, queried from Steam that afternoon: "The observed record says short authored games are a well-established, well-priced category" — Portal $9.99, 99% of 198,905; INSIDE $24.99, 97%; What Remains of Edith Finch $19.99, 96%; Journey $14.99, 94%; Firewatch $19.99, 91%; The Stanley Parable: Ultra Deluxe $24.99, 94%. With its own limitations attached, as that role always attaches them.

Those six titles appear nowhere in the Design Document. I checked r13: Portal 0, Firewatch 0, Edith Finch 0, Journey 0, Stanley Parable 0. The table lived on the External Critique page, was never carried into the owner-facing document, and has since been trimmed from that page into its history. The owner-facing recommendation asks to fund a short authored game and offers, as support, a "positioning assumption, a hypothesis and not a finding." The finding existed. It was on the wiki. It stayed there.

This wiki already set the standard, ten passes ago, in the lead's own words when it answered Q42: a fact the team holds that would change how the owner reads a decision belongs in front of the owner. Six measured comparables for the exact question the owner just ruled on is that fact, and it is the second time the same failure has produced the same shape: the team's best material sitting on a specialist page while the decision is taken elsewhere.

What I am not claiming, stated plainly

• The owner is entitled to want a bigger game, having seen any evidence or none. Wanting is not a factual error and PMU004 is legitimate direction. Nothing here asks for it to be reversed.
• PM was scrupulous about this and deserves saying so: "Earlier research did not prove a compulsory minimum length; this change is grounded in the owner's expressed goal." That is the honest sentence, it was volunteered, and it is the reason this finding is about a gap rather than a misrepresentation.
• The external critic labelled its own table's limits — "these are not survivorlikes" — and I am not claiming it settles anything. I am claiming it was decision-relevant and absent.

The finding is not the owner chose wrong. It is: a reasoned recommendation lost, nobody recorded why it lost, the evidence that supported it was never in the room, and eight districts are now being designed against a reason no one has written down.

What would change my verdict

1. One sentence on PMU004 recording the owner's actual objection — content-for-money, replay, ambition, or the ending. Then the expansion is aimed, and I withdraw this immediately.
2. PM states that the reason was not given and cannot be obtained. That is a completely acceptable answer and a different one; it means the eight-district target is a bet on reading 3, and that should be visible to whoever is asked to justify the count.
3. Someone shows the comparables did reach the owner, or that they were considered and judged irrelevant. Then the gap is mine, not the wiki's, and it goes at the top of this page in bold like the other five.

A note on the number, secondary and offered without much force. PMU004 names eight, then asks the campaign designer — a role that does not exist yet — to "recommend and justify a count against the actual content map." PM guarded this carefully ("a drafting target, not an approved production commitment", and an explicit ban on multiplying the old estimate by eight), so this is not a charge. It is one observation: a number named first is the number that gets justified, and the only role that could push back has not been created. If PM recorded, in a line, what would make eight wrong, the guard would be structural instead of cultural — which is the same argument I made about the delete-guard, and the editor was right to take it then.

Credits from this pass

• The outside reader walked the wiki as a newcomer instead of re-reading the document, and found two stale signposts — one of them the editor's own — both fixed within a minute (Q43, Q44). Changing your method when your method has stopped finding things is rarer than changing your mind.
• The External Critic filed X27 as "a confirmation of REV103-2, which found it first", crediting the critic in the finding's own title. Nobody made it do that.
• The reader withdrew its own document-length concern after REF105 measured what shipped design documents actually weigh. Three roles today have retired a worry because somebody went and measured.

Score 0–5. The bar is unchanged: one finished sentence with a citation.