INDUSTRY FIELD GUIDE
Surface coal sample and lot records for AI review
An observation from one portion of a stockpile cannot automatically be attributed to the whole lot.
NAICS 212114 · Surface Coal Mining
An observation from one portion of a stockpile cannot automatically be attributed to the whole lot. USGS publishes coal sample and quality data. A surface operation can use synthetic labels and quantities to test whether an assistant preserves the scope of a sampling record. USGS coal quality data
Invent two stockpile aliases and three sample labels. Give one sample a confirmed lot association and another only a date. Ask which sample supports the mock report and where attribution is unresolved. Keep sample basis, date and source distinct. Add a combined record that cannot be disaggregated without another source.
Do not share actual mine or material locations, quantities, client specifications, quality data, photographs of operations or control details. Use an independent reviewer and no site access.
Describe a fictional sampling exercise
Original Fiund exercise
One attributed sample cannot label a whole stockpile
The practice surface-coal ledger contains two stockpile aliases, three sample tags and a combined summary. No real coal quality or stockpile quantity is used.
| Record | Source | What it records |
|---|---|---|
| SL1 | Fictional sample tag S-A | Confirmed association to lot Cedar; observation U=4 practice units for this sample. |
| SL2 | Sample tags S-B and S-C | S-B shares S-A's date but has no lot association. S-C is confirmed to lot Birch. |
| SL3 | Combined summary | Reports U=6 for material combined from Cedar and Birch; no allocation or weighting is supplied. |
Question for the model
Which sample can be cited as evidence from Cedar, and what Cedar-wide or S-B claims remain unsupported?
Show the reference answer and checks
SL1 directly associates S-A and its observation of 4 practice units with Cedar. SL2 supplies only a shared date for S-B, so its lot remains unresolved. SL3 combines Cedar and Birch without allocation or weighting. It cannot establish Cedar's separate value, and S-A alone does not document representativeness of the entire lot. Keep sample-level attribution separate from whole-lot inference.
- Using a shared collection date to assign S-B to Cedar.
- Reporting the combined value or one sample observation as Cedar's established whole-lot value.
Fields to preserve
- Sample-to-lot evidence
- Identifies a confirmed association rather than a date-based guess.
- Sample scope
- Records which portion the observation actually describes.
- Combined-report composition
- Shows which lots contributed without inventing allocation.
- Representativeness evidence
- Makes the absence of a lot-wide inference rule explicit.
Test another case
Provide a verified S-B-to-Cedar correction but no sampling-design evidence. The answer should resolve the identity gap while still withholding a whole-stockpile quality claim.
Do not include real material quality, mine locations, quantities, customer specifications, operations photographs or control-system information. The exercise is an offline records task.
Discuss a contribution
Describe the sources you control, the task they could support and what has already been checked. Use the existing-data assessment for an archive or the collection brief for a new recording project. The contributor agreement defines Fiund’s licensing authority and compensation. A guide or fictional exercise does not establish an accepted collection or a sale.