OENRA
Reads a wine's chemistry and a person's palate on the same eight sensory dimensions to predict who will actually enjoy the bottle — and explains the match instead of returning a score.
A chemistry-aware wine intelligence system. OENRA turns analytical chemistry — GC-MS, LC-MS, FT-IR, routine panels — into a structured sensory profile of a wine across eight dimensions, each carrying a confidence that widens when the data is thin. A person's palate is modeled on the same eight axes, so a prediction isn't a universal 92-point score; it's an explained, per-dimension match between one bottle and one drinker, with the reasons and the uncertainty shown. For the trade, the same representation feeds batch-quality insight and blend exploration. It's a research preview: capabilities are experimental, and I make no accuracy or adoption claims.
The problem
A 92-point score was never about you. Critic ratings answer is this a good wine — a question about the bottle. The question a person actually has is will I like this one, which is a question about the bottle and the drinker. Those are different problems, and the wine world mostly answers the first and hopes it stands in for the second.
It doesn’t. Even expert critics don’t fully agree with each other, and crowd ratings track them more loosely still; in blind tastings, non-experts don’t enjoy more expensive wine more, and the stated price alone can raise reported pleasantness. A single number collapses a personal, multisensory experience into a scalar and then prints it on a shelf tag as if it were the same for everyone.
The thesis
A wine score answers a question you didn’t ask. The one you asked is whether you, specifically, will like this bottle — so model the person, not just the wine.
Measure. Model. Match.
OENRA puts the wine and the drinker into the same coordinate system. Analytical chemistry — routine panels and, where available, GC-MS, LC-MS, and FT-IR — is normalized into a structured representation of the wine. That representation is expressed across eight sensory dimensions — fruit, ripeness, acidity, tannin, body, oak, sweetness, savoriness — each estimate carrying a confidence that depends on how much data stands behind it. A person’s palate — their Taste DNA — is modeled on those same eight axes. The prediction is a comparison of the two, and it explains itself: which dimensions drive the match, which pull it down, and how sure the model is. You can operate the comparison in the Taste-DNA match explorer.
Why confidence is a first-class output
The eight dimensions are a chosen coordinate system, not a claim of biological ground truth — and the honest version of this shows its uncertainty instead of hiding it. A wine with rich analytical chemistry earns a narrow estimate; one identified only from its label gets a wide one. Surfacing that interval is the same discipline I bring to Littman and FlightReady: a high-stakes prediction should expose the tail and defer when the data is thin, not launder a guess into a confident-looking number.
The honest limit
Chemistry under-determines enjoyment. Instrumental measurement predicts a few composition-linked attributes moderately well and many aroma attributes weakly, and flavor is ultimately constructed by the brain from taste, smell, expectation, price, and context. So there is no defensible “we predict taste with X% accuracy” number, and OENRA doesn’t invent one. It models a real, useful signal — the drivers of an individual’s liking — and is explicit about where that signal runs out.
Status & the paper
Research preview. The wine representation, sensory model, and Taste-DNA preference model are experimental; chemistry ingestion and blend exploration are pilots. The full method — and the peer-reviewed literature it rests on — is written up as a working paper, from molecules to enjoyment, with the open questions it raises collected under research.