# Effort estimation study — your checklist

> Current status (2026-10-21): E07 reduced v3 is technically complete, with study acceptance pending; E01–E06 remain proposed. Every future cycle is enrolled under P2. The setup checklist below is historical; see the dated update at the end.

Current position: setup is complete on `feat/effort-estimation-trial`. Review the [study ledger and six proposals](./effort_estimation_trial.md), confirm the provisional anchor and choose the first candidate. No feature has been selected or implemented for the study. The agent maintains estimates, evidence, plans and the ledger.

## Finish setup

- [x] Approve revised Step 2 for direct vendored-workflow integration and projected-score checkpoints.
- [x] Approve Step 3 and authorize setup implementation.
- [ ] When setup is delivered, review the proposed candidates: are they useful, still wanted and accurately scoped? Flag anything already done or misunderstood.
- [ ] Review the provisional 1.0 reference: does it represent a complete slice you would comfortably approve for one cycle? Accept it or ask for clarification/revision.
- [ ] Choose the first candidate you want delivered. You do not need to fill size categories.

## Repeat for each selected feature

- [ ] Confirm the desired outcome and any important device, environment or recovery needs. Answer scope questions the agent cannot resolve from existing feedback.
- [ ] Review the feature’s plans and request changes or give each required approval: **Approved Step 1** if used, **Approved Step 2**, then **Approved Step 3**. Study setup approval does not approve these features.
- [ ] If the agent identifies acceptance checks requiring your access or device, perform those agreed checks and report what happened.
- [ ] Review the delivered result. Tell the agent about missing behavior or failed checks, distinguishing corrections from newly desired features.
- [ ] Give your comfortable-fit judgment: did this scope feel suitable for one cycle, too large, or smaller than expected? A brief reason is useful; you do not need to estimate hours.
- [ ] Choose the next candidate, replace one you no longer want, or pause the study. Ask the agent to resume from the ledger in a new session.

## After six completed outcomes

- [ ] Ask the agent to prepare the study review as a separate documentation cycle, and review/approve its plans.
- [ ] Review its conclusions and limitations, then decide whether to retain or revise the reference scale for a future batch.

You do not need to manage hashes, Git commits, estimate tables or evidence collection yourself. The agent handles those within the approved scopes. You also do not need to supply a second independent rating unless you choose to participate as a rater.

## Update — 2026-10-21

The opening status is historical: E07 was selected and its reduced v3 scope is technically complete; comfortable-fit judgment and acceptance remain pending. Six other candidates remain proposed. The maintainer has authorized provisional beta use and [standing enrollment](./effort_estimation_trial.md#continuing-enrollment--2026-10-21-protocol-p2) of every future cycle. No repeated study selection is needed. Feature approvals and individual acceptance still apply; six accepted outcomes are a review checkpoint and collection continues afterward.