---
name: quanterlab-progress-report
description: Regular status on QuanterLab research without the model inventing findings to fill the space. Use whenever a status, update, summary or progress report on studies is asked for, at the end of a working session, or when a walk finishes. Triggers on words like status, update, where are we, what happened, progress, report back.
---

# QuanterLab progress report

A report is a record of what changed, in the order a reader needs it. It is not the place where significance is discovered.

## What a report covers, in this order

1. **What was run since the last report.** Each study by name, the windows walked, the registrations made. Counts, not adjectives.
2. **What came back.** The logged numbers per window, book against benchmark, and the pooled figure if a walk completed. Excluded windows named with their reason. A number is quoted from the log, never from memory.
3. **What is still in flight.** Walks running, windows remaining, the expected time to the tally, anything blocked and on what.
4. **The next decision point.** The one choice the reader has to make now, with the two or three options and what each costs. Not a list of everything that could be done.

Anything outside these four belongs in the study's own record, not in the report.

## Verdicts reported as verdicts

A test that has a categorical outcome gets a categorical sentence: the harness passed or failed, the seal held or was refused, the consistency check passed or listed its failures, the window ran or was excluded. Do not soften a failure into "mostly worked" and do not round a pass up into "confirmed".

When a walk finishes, the report states the count of windows ahead of the benchmark, the count behind, and the pooled Sharpe with its trial count, and stops there. Interpretation is the paper's job, after the audit.

## Permission to say nothing happened

A period with nothing to report gets a report that says so: "No windows completed since the last report; the walk is at window 5 of 8; nothing is blocked." That sentence is complete and correct. The alternative is a model manufacturing a finding from a null result, or from a single window, which is the thing most worth keeping out of a research record.

A null result is reported as a null result, with its numbers. It is not reframed as a lesson, an insight, or a setup for the next idea.

## Language

- Short sentences. One number per sentence where a number is needed. No adjectives on results.
- Name studies by their registered names and windows by their anchors.
- Say "did not beat" rather than "lost to"; say "excluded" rather than "skipped"; say "refused" when the platform refused something and give its reason verbatim.
- Do not thank, apologise, or promise. Report.

## The last line

The last line of every report is the next decision, or the words "no decision needed until the tally lands", with the expected time.