Test & lab auction analysis methodology

Evidence hierarchy

1. 1. Seller-confirmed facts

   Explicit auction fields and seller statements, preserved with their source and observation date.

2. 2. Photo-derived observations

   Visible labels, modules, quantities, damage, and omissions—never upgraded into proof of function.

3. 3. Independent technical context

   Manufacturer and authoritative documentation used to understand the product family and dependencies.

4. 4. Market evidence

   Realized sales preferred; asking prices and related models remain labeled and discounted where used.

5. 5. LotCaliper assumptions

   Resale, logistics, testing, repair, risk, required profit, and ROI inputs kept separate from source facts.

6. 6. Unresolved unknowns

   Missing identifiers, configuration, condition, costs, and transfer rights remain visible and reduce confidence or prevent a bid limit.

Identity, configuration, and condition

Comparable-market evidence

Comparable evidence is normalized for exact model, configuration, condition, completeness, price type, and timing. Weak or non-realized evidence is not presented as a realized sale. Dealer and marketplace asking prices remain distinct from completed sales, and related-model evidence remains distinct from an exact match. These limitations reduce confidence and can narrow the usable resale estimate.

Market-evidence strength is separate from public auditability. Where source-use rights permit, an analysis can show individual comparable records. Otherwise it shows only truthful aggregate evidence metadata and explains the disclosure limit; limited public auditability does not by itself mean the underlying market evidence is weak.

Cost model and risk-adjusted maximum bid

Lot-specific costs can include a buyer premium, pickup, packing, freight, testing, refurbishment or repair, missing accessories, selling-platform and payment costs, and a reserve for resale friction and unresolved risk. Minimum profit and minimum ROI are separate return requirements, not costs. A cost is included only when it applies to the lot; an unknown charge stays visible rather than being treated as zero.

Each published page shows the lot-specific amounts and bases. If the premium is unknown, the ceiling is explicitly a hammer-bid limit and the buyer must account for any additional charge before bidding. The maximum is a reseller ceiling, not a forecast of the closing price.

Decision labels and confidence

Confidence is reported separately for identity, configuration, condition, market support, and costs. A strong model identification does not compensate for weak operating evidence or unknown logistics. Later bidding can move an opportunity above its recorded ceiling; the published label remains the decision made for the displayed snapshot.

High means direct evidence resolves material questions for that dimension. Medium levels indicate bounded estimates or verification needs; low levels indicate missing, contradictory, unbounded, or decision-dominant uncertainty. These labels measure evidence quality, not the likelihood of a profitable resale.

Numeric BUY and REVIEW decisions include one or two scenarios aimed at their most material downside, such as weaker resale, higher logistics, repair exposure, or a missing key component. A generic simulation is not substituted for lot-specific judgment.

How analyses are produced and who is accountable

LotCaliper is operated by Oleksii Udod, the accountable publisher for LotCaliper analyses. AI tools may assist with organizing and comparing source evidence, research synthesis, drafting, structured extraction or classification, consistency checks, and implementation or publication integration. The operator remains accountable for deciding what is published.

AI-generated output is not itself source evidence. Seller and listing facts, photo-derived observations, manufacturer or other authoritative technical context, market evidence, and LotCaliper assumptions remain separately identified. Unresolved facts stay unresolved rather than being filled from model output, and numeric resale and maximum-bid conclusions remain tied to structured evidence, costs, and assumptions.

Before publication, the operator reviews material decision inputs, including identity, configuration, condition, comparable-market treatment, cost and economic assumptions, and the recommendation and maximum-bid conclusion. Deterministic validators and tests help catch structural, mathematical, and semantic contradictions; they are consistency safeguards, not replacements for evidence judgment.

When the independent review gate applies, it is a separate source-first LotCaliper challenge pass. AI assistance may support that process. It is not an outside certification, manufacturer review, appraisal, or independent human domain-expert opinion.

Independent review and corrections

A separate source-first review pass is required for every BUY and for a REVIEW with conservative resale of at least $5,000. Lower-value REVIEW, SKIP, and INSUFFICIENT DATA analyses can be not required under the current gate. The pass starts from preserved source evidence and tries to challenge the proposed analysis before using its maximum bid as an anchor.

Some otherwise-not-required new analyses are selected by an auditable shadow sample for the same structured review. That selection expands quality-assurance coverage; it is not a stronger endorsement of a selected analysis.

It checks identity, configuration, condition, market relevance, conservative resale, logistics and repair costs, bid math, and the decision. Approved means every required area passed without a corrected review check. Approved with corrections means the review required one or more corrections before approval.

Not required means the current mandatory gate did not require a second review; it is not an approval. Legacy unreviewed means the page belongs to the frozen corpus published before the current gate and no completed independent review is recorded. PENDING and REJECTED high-stakes reviews do not pass the publication gate.

A completed LotCaliper review is a process and quality-control status, not an external certification, appraisal, guarantee of condition, resale result, or auction outcome. When stronger evidence or review findings change a published fact or conclusion, append-only correction records retain the original value, corrected value, reason, and whether valuation, maximum hammer, or decision changed.

Immutable decision snapshots

Every analysis preserves its original observation time, auction price or no-bid state, evidence, economics, and recommendation. Before the scheduled close it appears as an open-auction snapshot with a stale-data warning. After that time, the same URL automatically renders as historical. Exceptional cancellations, extensions, and relistings can be recorded explicitly without rewriting the original decision record.

Limitations

• LotCaliper does not inspect, test, calibrate, authenticate, or take custody of auction equipment.
• Resale ranges and maximum bids are models, not appraisals, guarantees, or individualized financial advice.
• Buyers must verify current auction status, eligibility, taxes, fees, removal terms, shipping, configuration, condition, and legal or regulatory requirements.
• Clinical suitability, radiation safety, calibration status, software transferability, and regulatory compliance are never inferred from auction photos.

See the method applied