chore(site): re-record the demo replay with the grade formula 2026.3 card model #193

Merged
ryangr0 merged 3 commits from ryangr0/ploeg-card-grade-review-rounds into development 2026-10-03 15:13:33 +00:00 AGit
Owner

Vloer's card model now describes grade formula 2026.3 and the inputs a
grade missed, and the demo's illustrative 2026.2 grades carry a null
reworkRounds, so the hosted replay no longer matched what Vloer serves.
Re-recorded with mise run demo-record: the card model hash, the replay
hash, the commit, reworkRounds on the demo grades and recorded test
timings change.

Three commits: Ploeg grade formula 2026.3 + proposed ADR-0061, Vloer display, separate chore(site) replay re-record. Review loses 1 per rework round (a head where a human requested changes); approvals and comments cost nothing. A subgrade with a missing input is capped at 9 and the card earns no label; inputs.missing lists them. No snapshot or migration: grades stay computed on read, and run-cards.md no longer promises old grades never change. Owner questions: unreviewed merges still score review 10; per-round vs per-review rework counting; whether the cap should also cap overall.

Verified with mise run verify on the pinned toolchain (all gates passed).

Ticket: https://vikunja.webgrip.dev/tasks/1751

🤖 Generated with Claude Code

Vloer's card model now describes grade formula 2026.3 and the inputs a grade missed, and the demo's illustrative 2026.2 grades carry a null reworkRounds, so the hosted replay no longer matched what Vloer serves. Re-recorded with mise run demo-record: the card model hash, the replay hash, the commit, reworkRounds on the demo grades and recorded test timings change. Three commits: Ploeg grade formula 2026.3 + proposed ADR-0061, Vloer display, separate chore(site) replay re-record. Review loses 1 per rework round (a head where a human requested changes); approvals and comments cost nothing. A subgrade with a missing input is capped at 9 and the card earns no label; inputs.missing lists them. No snapshot or migration: grades stay computed on read, and run-cards.md no longer promises old grades never change. Owner questions: unreviewed merges still score review 10; per-round vs per-review rework counting; whether the cap should also cap overall. Verified with `mise run verify` on the pinned toolchain (all gates passed). Ticket: https://vikunja.webgrip.dev/tasks/1751 🤖 Generated with [Claude Code](https://claude.com/claude-code)
Formula 2026.2 took 1 off review per changes-requested review and 0.5 per
human review round beyond the first, including approved and comment-only
rounds, so a more careful review lowered the card. A cost that was not
reported, nothing authorized, or a board without gates left budgetShare or
defectBounces null and cost nothing, so a card with unknown delivery facts
could score and label as well as one with known clean facts. The concept
page promised that a formula change never re-grades old cards, while Ploeg
computes every grade on read under the current formula.

Formula 2026.3 (ADR-0061, proposed) takes 1 off review per rework round, a
distinct (play, head commit) on which a human requested changes; approvals,
comments and extra rounds cost nothing. inputs.missing names each input
Ploeg can collect but has no fact for on the card; such a subgrade is at
most 9 and the card earns no label. inputs.review.reworkRounds is new. The
grade stays computed on read with no stored snapshot, and the docs now say
so: the card names the formula, and a formula change re-grades every card
read after it. The operator-api schema, contracts README, ADR index, ADR-0050
and ADR-0052 cross-references, the Run Cards page and the glossary follow.

VIK-1751

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Ploeg now grades under formula 2026.3, which this Vloer could not describe:
the Grade tab fell back to "cannot describe", the condition rows only read
"No confirmed crack" for 2026.2, and the new reworkRounds and missing
inputs were dropped by the card proxy.

The proxy passes review.reworkRounds (null when an older Ploeg did not send
it) and inputs.missing (absent when not sent). The Grade tab describes
2026.3, lists rework rounds, marks each missing input as "Missing for this
card" with its cap, and shows an Evidence row, complete or incomplete, only
when Ploeg sent a missing list. Condition treats 2026.2 and 2026.3 alike.
The demo's illustrative grades stay on formula 2026.2, so their rework
rounds are null.

VIK-1751

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
chore(site): re-record the demo replay with the grade formula 2026.3 card model
Some checks failed
[Workflow] On Pull Request / release-policy (pull_request) Successful in 1m2s
[Workflow] On Pull Request / checks (pull_request) Has been cancelled
[Workflow] On Pull Request / warnings (pull_request) Has been cancelled
20be1721c2
Vloer's card model now describes grade formula 2026.3 and the inputs a
grade missed, and the demo's illustrative 2026.2 grades carry a null
reworkRounds, so the hosted replay no longer matched what Vloer serves.
Re-recorded with mise run demo-record: the card model hash, the replay
hash, the commit, reworkRounds on the demo grades and recorded test
timings change.

VIK-1751

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
ryangr0 merged commit 366484b566 into development 2026-10-03 15:13:33 +00:00
Sign in to join this conversation.
No reviewers
No labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
webgrip/unfold!193
No description provided.