I3371 — Economics summary decides a registered suite's primary outcome by its registered test, reports secondary differences without a verdict word, and refuses typed figures about them
For a registered suite compute per task and arm the accepted share, summed session wall time and reported spend (tokens, and the harness's projection labelled as such); decide the primary outcome with I3366's exact test at the registered level; report secondary outcomes as mean paired differences with task-clustered intervals and no verdict word; state the effective sample as tasks; report status running until the registered sample is complete; report per-stratum results as preliminary and human interventions and CI reruns as not observed. Render the section in the generated report and the site page after I3356's scorecard section, and extend the economics vocabulary so a typed acceptance, wall-time or intervention figure is refused.
BLOCKED wave 5 · p1 · implementation profile · parallel safe
Blocked. This issue cannot start until I3366, I3367, I3356, I3133 are done. The status is derived from that, not declared.
Objective
For a registered suite compute per task and arm the accepted share, summed session wall time and reported spend (tokens, and the harness's projection labelled as such); decide the primary outcome with I3366's exact test at the registered level; report secondary outcomes as mean paired differences with task-clustered intervals and no verdict word; state the effective sample as tasks; report status running until the registered sample is complete; report per-stratum results as preliminary and human interventions and CI reruns as not observed. Render the section in the generated report and the site page after I3356's scorecard section, and extend the economics vocabulary so a typed acceptance, wall-time or intervention figure is refused.
Why
T2 of the owner's report of 2026-10-19 asks for the comparison to be computed, not described. The critique asked for one verdict on the primary outcome only, the sample stated as tasks, and no interim verdict; it depends on I3356 because both edit the economics summary, its schema and the economics template.
Current state
apps/majordomus-cli/src/economics/mod.rs:1809 derives completion_rate per arm; grep -n duration_ms mod.rs returns nothing though model.rs:687 records it; claims.rs:41-44 VOCABULARY holds no acceptance or wall-time word.
Desired state
economics summary, its schema, docs/generated/economics.* and the site carry the registered-suite section; economics check refuses a typed figure about these outcomes.
Scope
- apps/majordomus-cli/src/economics/mod.rs
- apps/majordomus-cli/src/economics/model.rs
- apps/majordomus-cli/src/economics/report.rs
- apps/majordomus-cli/src/economics/claims.rs
- apps/majordomus-cli/tests/economics.rs
- share/schemas/generated/economics-summary.schema.json
- docs/generated/economics.json
- docs/generated/economics.md
- docs/generated/economics.yaml
- site/data/generated/economics.json
- site/templates/economics.html
- docs/ECONOMICS.md
- site/content/docs/economics.md
Out of scope
- The acceptance definition and pairing (I3133)
- Live runs (I3372)
Dependencies
- I3366READY The economics calculator decides a binary paired outcome by an exact task-level test, intervals a continuous paired difference by task-clustered bootstrap, and turns either into one verdict
- I3367BLOCKED A suite can be registered: strata, outcomes, acceptance rule, analysis, MDE, a seeded interleaved run order and the stopping rule are declared data the load validates and the runner honours
- I3356BLOCKED The site's economics page publishes the scorecard from the generated summary, failed and inconclusive tests and every loss in the same rows as passed ones, each linking its records
- I3133BLOCKED The economics benchmark measures acceptance: how many sessions declared done while the task's hidden tests fail, with and without the finish contract
What waits on this
- I3372BLOCKED The registered corpus is run in both arms, its records are committed, and its verdict is published with its losses and its origin
- I3385BLOCKED A registered suite's primary verdict can be held to a non-inferiority guard, so a speed-up counts only when accepted share is not worse by more than a registered margin
Acceptance criteria
- Fixture runs where the treatment accepts more tasks give a favourable primary verdict only when the exact test rejects
- Fixture runs below the registered sample read running with no verdict word
- Secondary outcomes print a difference and interval and no verdict word
- Human interventions and CI reruns read not observed
- economics check refuses a hand-typed acceptance or wall-time figure in prose
- CLI, JSON, MCP and the generated report agree
Validation
- cargo test --manifest-path apps/majordomus-cli/Cargo.toml --test economics
- bash test/run.sh 430_economics
- bash test/run.sh 431_economics_claims_gated
- majordomus economics check
- majordomus generate --check
Evidence required
- test_registered_suite_primary_verdict
- test_corpus_figures_refused_in_prose
Evidence
None recorded. Every token above needs a command or an artifact behind it before this issue can be completed; narrative is refused.
Timeline
- started
- —
- verified
- —
- completed
- —
Those three fields, the evidence above and the state of the dependencies are all the status is made of. There is no status field to disagree with them.
Canonical record: .ai/repo/project/issues/I3371.yaml. Read it back with majordomus plan show I3371.