Grading Philosophy

Structured evidence. Consistent criteria. Final human accountability.

The NOSTALGI-Ai promise

We will pursue innovation without abandoning responsibility. We will use artificial intelligence to improve human decision-making, not hide it. We will treat every artifact with respect regardless of its financial value.

A grade is not a reward, a punishment, or a price prediction. It is our documented professional conclusion about the condition of an artifact under the current NOSTALGI-Ai Grading Constitution.

The four condition factors

Our grading framework begins with four core areas. The detailed scoring rules and tolerances are still being calibrated before public submissions open.

Centering

Front and back alignment, border relationships, and image placement.

Corners

Sharpness, wear, compression, whitening, and structural integrity.

Edges

Chipping, fraying, rough cuts, discoloration, and localized damage.

Surface

Scratches, print lines, indentations, stains, gloss, and manufacturing defects.

Evidence before assumption

Grade what is observable. Use approved imaging evidence, measurements, tools, grading standards, and documented procedures. Do not invent defects, ignore visible evidence, or alter the conclusion because of the customer, card value, manufacturer, rarity, or market price.

How a decision is supported

1. Capture — Controlled imaging is used to create consistent documentation.
2. Observe — Relevant condition evidence is organized across the four factors.
3. Compare — The observations are evaluated against the current grading criteria.
4. Verify — A trained human reviewer independently examines the evidence and remains responsible for the final decision.
5. Escalate when needed — Proposed adjustments can require approval by an authorized second verifier or management grader.
6. Record — The planned condition record connects the decision to documented observations and approval history.

Calibration before scale

NOSTALGI-Ai is currently validating its procedures through an internal 100-card calibration phase. This work is intended to test repeatability, identify ambiguous criteria, and improve reviewer training before public volume.

We are not publishing accuracy percentages or consistency claims before there is enough real evidence to support them.

What will be published before launch

• The public grading scale and condition definitions.
• Submission, packaging, custody, and handling requirements.
• Pricing and expected turnaround times.
• Regrade, error, damage, insurance, and liability policies.
• The final contents of the customer condition record.

Current status

Public grading submissions are not open. The framework, training procedures, imaging workflow, and scoring criteria remain in active development and calibration.