Evidence Viewer
Decode probable production dates, compare runs, browse photographic evidence and help build a better dataset across Ex Nihilo fragrances.
Reported → Photo verified → Confirmed, based on independent evidence.
Optional longevity and projection data can be attached to each piece of evidence.
Records are scored based on photo, source, purchase details, confirmations and moderation.
Fast model interpretation with evidence status kept separate from authenticity.
| Fragrance | Batch | Date | X | Confidence | Research grade | Quality | Evidence | Performance |
|---|
Current, archived, special-edition and researched regional exclusives. A fragrance can appear here even when no batch has been documented yet.
Automatic analysis of the current dataset. These are research signals, not manufacturer-confirmed rules.
Multilingual batch discovery across marketplaces, perfume communities and official/reference sources. A public mention is a lead, not automatic proof of authenticity.
Not all evidence has the same calibration value. Wave 3 ranks evidence types by what they can independently prove: batch existence, fragrance identity, exact date and source independence.
External dates are compared with the ordinal decoder. Exact matches strengthen the model, but remain community/market evidence rather than manufacturer confirmation.
Mirrored listings are clustered so copied pages do not inflate confidence as independent confirmations.
Codes with prefix or catalogue contamination are blocked from normal decoding until resolved.
These stay outside the valid batch database and are used only by the anomaly/authenticity guard.
Subjective owner reports can reveal formulation or storage differences, but they do not prove reformulation. V8.3.3 stores them separately from production-date evidence.
Next searches are ranked by how much they can improve prefix identity, code transcription or cross-source confidence.
Compare alternative interpretations against the observed dataset. Scores are heuristic research-fit indicators, not official validation.
Wave 3 adds Outcast Blue Extrait OU13063 as a single-source exact-code lead and introduces an Evidence Strategy Lab. Authenticity cards, direct label photos and independent source clusters are now ranked by how strongly they can calibrate the decoder.
Resolved mappings and unresolved hypotheses are kept separate. Candidate mappings never auto-detect a fragrance until photo-verified or confirmed.
Photo leads, owner leads, retailer leads and conflicting identifiers stay visible until the batch can be read and independently checked.
| Fragrance | Lead | Code / prefix | Size | Priority | Research note |
|---|
This page tracks which product prefixes have actually been observed and how strong the evidence is. Prefixes are not inferred from fragrance names unless evidence exists.
| Fragrance | Prefix | Status | Evidence | Collection | Action |
|---|
Choose a fragrance to visualize its known production-code sequence and batch families.
Known records show evidence and confidence. A valid but unknown code can also be compared as an Experimental model-only decode.
Best submissions include: a clear batch photo, correct fragrance/concentration, bottle size, approximate purchase period, country and public source.