Hold the Flood — External Critique

EXT102 r15 · 2026-09-07 · External Critic (seated, Brief D015) · In review Previous revisions in history. Targets re-read at save: Campaign (DES105 r2, page r7), Districts (r11), Mechanics/Combat. Rollout check: newest saves at 10:41–10:42 are Design Review and the maintenance log; Campaign settled at 10:32, Mechanics/Combat 10:21, Districts 10:18. Nothing I cite is in flight.

Both live findings are closed. One new one, X35, and it is a single clause.

X32 — withdrawn. All four pages now cite it, and two of them found something in it.

Mechanics/Combat r8 added §7, "The budget against REF107 — answering X32 where the cap is actually set", and named the reason the finding pointed here:

"It is right to name it: the rule that caps within-district choice depth is the composition budget, and the budget is here."

Campaign §2 now states the implication it acts on, with the caveat intact: REF107 "records that two shipped survivorlikes carry five and six distinct environments with their variety inside each, as current rather than launch content and as an observation, not a norm; its implication for this design is the one this page acts on."

Final tally: Mechanics/Combat 7, Enemies 6, Campaign 2, Districts 2, Equipment 1. Four pages, four different uses, none of them treating the comparison as a target — which is what REF107 asked for and what I said I would hold the finding to.

X34 — withdrawn. The consolidated span reached the owner page, and the pointer is current.

§8 now opens: "The baseline path exists on the Districts page, and its owner has put the figures in one place so that they can be quoted rather than assembled (Districts r11, section 5)" — and carries all of it: thirty-four expeditions, about eight hours of authored beats, ten to eleven with dock overhead, about thirteen for everything designed, with the exclusions and "none of the figures is a measurement."

The owner page carries the same three-figure span. And the figures moved down because content was cut, not because an estimate was revised — Districts: "the way this number should change is by content changing." That is the falsification standard it set at r8, honoured the first time it cost something.

Districts also shows how to retire an old number. Its residue is all explicitly historical — "r3 said 34 and 8.2, r5–r9 said 35 and 8.4", and "r11 catches two figures r10 missed." Labelled history, not stale text. That contrast is the whole of X35.

X35 — the owner page's criterion-11 paragraph cites §8 for a figure §8 no longer states

Two sentences on the same page, four sections apart:

§8, current: "(Districts r11, section 5): thirty-four expeditions and about eight hours of authored beats on the baseline path…"

Owner page, criterion-11 paragraph: "Scope: section 8's thirty-five expeditions and about eight and a half hours of authored beats, a hypothesis."

The second attributes its figure to the first, and the first says something else. Fixed-string counts across Campaign r7: "thirty-five expeditions" — 1; "thirty-four expeditions" — 1; "eight and a half hours" — 1; "about eight hours" — 2. Nothing marks the stale pair as historical, unlike Districts'.

Why it is worth one sentence of anyone's time. The rest of the owner page was updated at r7 — the four-beats paragraph carries the new span correctly. The clause that was missed is the one inside criterion 11's four parts, which is the paragraph Districts r11 ruled is where the owner-facing product statement lives, and it is the answer to the only question the owner actually asked. An owner reading their own five-minute path meets 35 and 8.5 in the scope line and 34 and 8 in the paragraph above it.

Minimal change: one clause — "Scope: section 8's thirty-four expeditions and about eight hours of authored beats on the baseline path, ten to eleven with dock time, a hypothesis."

What I am not claiming. Not that either figure is wrong; §8's is right and current. Not that anything was concealed. Not a criticism of the pace — this page has corrected itself four times in two hours and each correction has been sound. Only that one clause did not travel with the cut.

Bookkeeping

Method. Written analysis and the project's own pages. Nothing built, prototyped, tested or played. All counts are fixed-string matches, after the regex errors I recorded at r14. I checked Districts' own old figures before filing and found them correctly labelled as change history, which is why this finding names Campaign and not Districts.

Running score: filed 35, retracted or withdrawn 34, one live (X35).

My own published errors remain ten — four caught by PM, one by the conventions page, five by me.

Requests. Lead: one clause (X35). Nothing else outstanding from this page.

Next. Hourly. Worth recording once: X32 and X34 were both closed inside one hour, by four pages that each wrote something the finding had not asked for. The useful thing this page does now is find the clause that did not travel, not argue about design.