A record of what we learn when the format meets real people, and of what we changed because of it. Read the first paragraph before the lessons: a rehearsal is not a study. One session, a small group of educators standing in for teams, our own people watching our own format. We publish it because a format nobody can see being corrected is a format nobody can check — and it is why this page never says the evidence shows anything.
How to read a lesson
Each lesson has four parts: what we observed, what we decided, what changed because of it, and where that decision lives. A lesson can honestly be observed and not yet decided; when that is the case it stays on this page in that state, rather than being held back until it looks finished.
The rehearsal — 2026-08-18
One session, a small group of educators standing in for teams, watched by the people who designed the format. Field experience, not a finding.
What we observed — In rehearsal we observed that briefs that a generic answer from the machine could not win still left teams without enough footing to start.
What we decided — We decided a brief may carry a possible way through the problem — never scored, written for that brief alone, and thinning as the bands get older. The proposal was accepted only together with those conditions.
What changed because of it — The authoring guidance gained the conditions, and the judging instrument gained a rule of its own so the route never reaches a judge's hands.
Recorded in — ssot/desafios/00-meta/rehearsal-2026-08-18-adjudication.md · ssot/desafios/00-meta/rulings.md
What we observed — In rehearsal we observed that teams began building before deciding what would exist at the end.
What we decided — We decided the team commits to its own output format at the divergence moment, and that any menu offered stays open. One clause of the proposal was rejected: several renderings of a single concept never count as several different approaches.
What changed because of it — The gate gained a criterion and an open-menu wording; the judging instrument was deliberately left untouched, so the commitment stays a scaffold rather than becoming a scored dimension.
Recorded in — ssot/desafios/00-meta/rehearsal-2026-08-18-adjudication.md · ssot/desafios/00-meta/rulings.md
What we observed — In rehearsal we observed that a brief could be designed, and solved, without touching a single named learning objective.
What we decided — We decided the authoring record must name the objectives the brief adds, and that review returns any brief which can be completed without engaging them.
What changed because of it — The authoring record gained the field, and the designer's sign-off gained the check that makes it unavoidable.
Recorded in — ssot/desafios/00-meta/rehearsal-2026-08-18-adjudication.md · ssot/desafios/00-meta/rulings.md
What we observed — In rehearsal we observed that teams crossed the work as one undifferentiated mass of using the machine, and checking sources and outputs was the first practice to disappear.
What we decided — Carlos challenged the first landing of this lesson and it was argued again from the beginning. We decided the declaration of a process becomes a gate of its own, at every band including the youngest; that a set of frameworks is taught in plural with your own way always valid; and — overriding the re-argument twice — that the deliberateness of the choice is read while fidelity to it never is.
What changed because of it — A gate was added after the divergence moment, the taught set was curated and labelled, and the reading of deliberateness entered an existing row rather than becoming a new one.
Recorded in — ssot/desafios/00-meta/rehearsal-2026-08-18-adjudication.md · ssot/desafios/00-meta/finding-4-readjudication.md · ssot/desafios/00-meta/rulings.md
What we observed — In rehearsal we observed that solutions drifted toward the machine's average answer wherever nothing asked for the team's own material.
What we decided — We decided this is a requirement of the brief packet, read through rows the judging instrument already has. No new scored surface, and no change of weight.
What changed because of it — The packet gained the requirement, and two existing rows gained it as a source of evidence.
Recorded in — ssot/desafios/00-meta/rehearsal-2026-08-18-adjudication.md · ssot/desafios/00-meta/rulings.md
What we observed — In rehearsal we observed that reviewing and auditing what the machine produced happened only where the brief itself asked for it.
What we decided — We decided the gap was declarative rather than evaluative: the instrument already read this work, and the brief did not say so.
What changed because of it — The brief itself now declares that reviewing, auditing and adjusting the machine's output is a central and evaluated part of the work; the judging instrument was not changed, because it was never the problem.
Recorded in — ssot/desafios/00-meta/rehearsal-2026-08-18-adjudication.md · ssot/desafios/00-meta/rulings.md
What we did not change, and why
The rehearsal changed how briefs are authored and how a team’s process is made visible. It changed nothing about the rows or the weights of the judging instrument, and one part of one proposal was rejected outright. That is deliberate: the rehearsal observed authoring and visibility failures, not judging failures, and a lesson is not a licence to reopen whatever happens to be nearby.
Future runs, rehearsals and consultations
Each future occasion becomes a block of its own, with its date, what kind of occasion it was, an honest description of what it could and could not show, and its lessons in the same four parts. Blocks are appended, never rewritten. A lesson whose disposition later changes gains a dated amendment underneath it and keeps its original text.
A consultation — when the subject is another party’s practice — takes this shape when it appears: the date; who was consulted, by role and number, never by name without documented consent; what was asked; what was observed, in the same observed register as above; and what was ruled, built or declined because of it. Each future occasion, of any kind, carries its own clearance before it appears here — clearance is per record, never per kind. No consultation has been recorded yet, so none appears.
How this page is maintained
It is generated from the adjudication records themselves, so a lesson cannot drift from what was actually ruled, and a new ruling reaches this page by regeneration.
Generated from the Challenges canon at version 0.20.0. Cleared for publication in the observed register by ruling D39 (2026-08-30), for this record alone.