PYTHON · NUMPY

Reporting what you cannot trust

Why report untrusted volume separately instead of folding it into unknown?

Write a sensing disclosure that states resolved, unknown and untrusted volume, and defend it against the objection that admitting uncertainty weakens the claim.

01Challenge

You have to hand a season of mesh data to someone who will act on it: a verifier, an insurer or a grower. Write the one paragraph that goes with the data, and make it survive a reader who wants to know what you could not see.

Predict firstWhy report untrusted volume separately instead of folding it into unknown?

02Model

There is a reason systems report coverage and not uncertainty, and it is not ignorance. A number with an admitted gap looks weaker than a number without one, and the reader is not always sophisticated enough to prefer the honest version. The answer is not to hide the gap and it is not to lead with it. It is to report all three quantities as one statement, because together they say something no single number can. Resolved volume says what was measured. Unknown volume says what no sensor reached. Untrusted volume says what a sensor reached while past its calibration interval. A reader who gets all three can price the claim. A reader who gets only the first has to trust it, and trust is what everybody in this chain is trying to stop relying on. The disclosure is stronger than the bare number precisely because it can be checked.

03Guided practice

Compute the three volumes from the mesh record, then write the disclosure. The rule for the prose is that every number a reader could challenge must be traceable to something in the record, and any quantity you could not determine is named rather than omitted.

04Feedback

PASS WHEN You wrote a claim a reader can check rather than one they have to trust. Notice the last paragraph: naming the quantity you did not measure is what makes the rest of it credible.

If your run fails, check:

trusted and untrusted are lists or sets of node ids. Use days_since_cal against RECAL_INTERVAL_DAYS.
The three volumes must sum to FIELD_VOLUME. If they do not, unknown is probably being computed from the trusted count instead of from the whole mesh.
Four nodes in this record are past interval. If you have a different number, check the season-end date.

05Retrieve & space

A sensing disclosure states resolved, unknown and untrusted volume, not coverage alone.
A number a reader can check is stronger than a number they must trust.
Naming the quantity you could not measure is part of the disclosure, not an admission against it.

06Mastery & project

Answer the objection directly: a colleague says publishing the untrusted fraction will cost you the contract. Give the argument that it wins the contract instead.

Write the disclosure paragraph for a real dataset you have produced or used. Include the quantities you could not determine. Then give it to someone who was not involved and ask them what they still cannot tell.

GO FURTHER · THE LIVE PLAYGROUND

The report this course serves ↗ · TR-2026-44 defines the unknown fraction. This course adds the band it does not yet compute.

← What recalibration is worth Back to course →

PYTHON · NUMPY

Reporting what you cannot trust

The mesh record for one season is given, including calibration dates. Produce the three volumes and the disclosure text.

You read

the arrays and values already in scope

You change

the code you write in each cell

Fixed

the dataset and the checks that grade you