Official
Current government, airline, manufacturer, or product documentation. Official defines the published boundary; it does not prove an exact product pair works well.
How we verify
Evidence before pages.
A useful database can say “we don’t know yet.” This method prevents source authority, product testing, and community experience from being blended into one misleading score.
Evidence levels
Five labels, five different claims.
Tested
An editor followed the published bench and real-use protocol with identified machine, firmware, product, batch, and context.
Community verified
At least three complete, independent public or submitted reports identify the exact pair and agree under comparable conditions. Conflicting reports remain visible.
Single report
One traceable public or submitted report identifies the exact pair with enough context to be useful, but not enough to generalize.
Estimated
A material, dimension, or rule-derived starting point. Estimates are marked, reviewed, and kept out of search by default.
Hard rules outrank averages
A published dimension limit or excluded material cannot be erased by a high blended score.
Rules have versions
Sources carry access dates. Policy, firmware, formula, and model changes send affected results back to review.
Reports are moderated
Public posts and user submissions do not change conclusions until checked for exact model and SKU, context, duplication, conflicts, and disclosure. Network evidence is never labeled as a PawExact test.
Publication gate
Why only 13 result pages are currently indexable.
- 1
Entities are complete. Exact model, SKU, material, and current source are known.
- 2
The result is specific. It contains pair-specific reasons rather than a reusable paragraph.
- 3
Evidence is reviewed. Sufficient independent exact-pair reports support the wording, with conflicts and official boundaries still visible.
- 4
Critical rules are current. No expired or unresolved source controls the answer.
Add real evidence