OBJECTIVE: Test whether a fictional service report distinguishes completed durations from unfinished waits. Use paper records or an authorized isolated test environment with invented cases and a test clock. Do not delay real requests or alter live service records.
1. Build a Fixed Observation
Create four requests completed within one declared reporting window. Their elapsed durations are two, three, four and five test minutes. Their arithmetic mean is 3.5 minutes.
Add request E, still open at the observation time and already forty minutes old. The fixture now contains four completed requests and one open request. It does not establish E’s eventual completion time.
2. Inspect the Headline
The completed-duration mean should remain 3.5 minutes and be labeled as applying to those four completed requests. A claim that everyone is served within that time fails the exercise.
Show the open count separately, with E’s age tied to the observation time. Check the exported report as well as the dashboard: neither should quietly drop E or describe its forty minutes as a final duration.
3. Advance Only the Clock
Advance the test clock by ten minutes, without adding completions or moving the declared reporting window. E is now fifty minutes old. The four completed durations and their mean have not changed.
Verify that the open-age view advances while the completed summary remains correctly scoped. A page refresh must not reset E’s starting event or make the aging request disappear.
4. Record a Real Test Completion
In the fixture, complete E at a duration of fifty-two minutes. Define a reporting window containing all five completions. Their mean is now thirteen minutes: sixty-five total minutes divided by five requests.
Completion criterion: E remains visible before completion without an invented ending, then joins the completed population under the declared rule. A reviewer can reconstruct the counts, time references and arithmetic from the synthetic records.
Read Completed Is Not Everyone and the fictional Wait Outside the Average for context.