Sensium one-way exam hubs 2026: 12 grapes the planting graph under-weights

Study 1a ranked mutual confusion edges by planting weight. Study 1b mapped neighborhoods around the 15 most-planted hubs. Both cuts are honest — and both systematically under-weight exam grapes whose catalog edges point out to famous look-alikes that do not always point back.

Headline finding: Nebbiolo, Carmenère, Corvina, Blaufränkisch, Pinotage, and several quiet whites sit outside the planting-mutual spine even though every serious tasting paper still tests them. Their dossiers carry four confusion targets that are mostly or entirely one-way: the hub lists Tempranillo / Pinot / Syrah / Sangiovese, but those high-planted grapes do not always list the hub among their four chips. The product still teaches the edge from the hub’s page; Study 1a’s top-25 table will not elevate it.

This Study 1c publishes 12 one-way-heavy exam hubs (planting ranks 21–70, ≥3 one-way neighbors in the four-chip cap) so syllabus priority and graph priority stop being confused. Still catalog-graph research, not population miss-rates (Study 2 / topic `137` when volume clears).

Companions: Study 1a — top 25 pairs · Study 1b — neighborhoods.

Methodology (read this before citing)

What this is not: a claim these are the grapes candidates miss most in live Train/Blind telemetry. It is a graph-structure finding: planting-weighted mutual rank and exam syllabus rank diverge, and the divergence is measurable in the catalog.

Why one-way edges exist (and why that is fine)

The editorial graph caps each dossier at four high-signal chips so Compare and mobile UI stay teachable. Cabernet’s four slots fill with Merlot, Syrah, Malbec, Cabernet Franc — all mutual, all high-planted. There is no spare chip for every grape that might confuse with Cabernet in a flight. Nebbiolo still needs to warn “don’t call Pinot” — so Nebbiolo → Pinot is stored on Nebbiolo even when Pinot’s four chips are already full of Sangiovese, Pinot Gris, Cinsault, and Barbera.

Study 1a answers: which edges does the product elevate globally? Study 1c answers: which exam hubs must you open on purpose because the global table will not?

The 12 hubs (curated from the filter)

→ = one-way from hub; M = mutual. Separators trimmed to the first catalog cue.

Pale / high-tannin / Italian–Balkan reds

Pepper / smoke / New World identity traps

Southern Italian warmth + Montepulciano

Quiet / name-trap whites

Three takeaways

1. Syllabus ≠ planting-mutual rank

If you only drill Study 1a’s top 25, you can still walk into Nebbiolo, Pinotage, Carmenère, and Aligoté cold. Schedule these hubs on purpose after the mutual spine is green.

2. Open the hub page, not only the famous target

One-way means Compare from Pinot will not always suggest Nebbiolo. Open Nebbiolo (or Train focus on that grape) so the four chips appear. The edge lives on the underdog’s dossier.

3. Parent / name traps are catalog-first separators

Pinotage ≠ Pinot / Cinsault and Aligoté ≠ Chardonnay are stored as explicit identity separators. Treat them as stop rules before aroma poetry — the graph is teaching label literacy, not only structure.

How to drill in three weeks

1. Week 1 — pale walls. Nebbiolo, Barbera, Corvina, Baga. Force tannin-vs-acid and rose-tar vs sour-cherry before place fantasy.
2. Week 2 — pepper / smoke / New World. Blaufränkisch, Pinotage, Carmenère, Zinfandel. Keep rotundone open.
3. Week 3 — south + quiet whites. Nero d'Avola, Montepulciano, Aligoté, Cortese. One neutral-white flight + one southern warmth flight.

Then return to Study 1a ranks 1–10 so the mutual spine stays automatic.

A useful study habit: after any wrong call on a famous grape (Pinot, Syrah, Tempranillo), ask whether an underdog hub should have been on the table — Nebbiolo, Pinotage, Carmenère, Aligoté. Open that hub’s dossier chips before you invent a second famous name. Many candidates only reverse-search from the planting spine and never see the one-way warning that would have saved the call.

Cross-check place when the miss is regional rather than cultivar: Italian, French, and Spanish hubs. Cross-check structure with tannin, acidity, body. The one-way list is a navigation tool — it tells you which dossiers to open; the deep dives teach the separator.

What we will publish next (Study 2)

When anonymized Train/Blind wrong-answer volume clears a documented n threshold, Study 2 will publish miss-rate ranks — the empirical test of whether one-way exam hubs are also population miss hubs. Until then, cite this page as catalog one-way structure, not live accuracy.

Frequently asked questions

Why didn’t these appear in the top 25 confusion graph?

Study 1a requires mutual edges first, then lower combined planting rank. One-way hubs fail the mutual gate even when they are exam gold.

Is “one-way” a data bug?

Usually no — it is the four-chip cap on high-planted grapes. Re-run the export after catalog edits; if a hub becomes mutual, it may graduate into Study 1a/1b cuts.

How is this different from the neighborhoods post?

Study 1b starts from planting ranks 1–15 (almost all mutual). This Study 1c starts from ranks 21–70 with one-way-heavy neighborhoods the global table under-weights.

Can I cite this in a class?

Yes — attribute “Sensium one-way exam hubs, 2026-07-09” and link this URL. Do not call them “most-missed” unless citing a future Study 2.

How do I reproduce the full filter list?

`node scripts/data/export_one_way_exam_hubs.mjs --min-pop 21 --max-pop 70 --min-one-way 3` — returns more than the curated 12; the table above is the teaching cut.

Open Nebbiolo in Compare this week, walk its four one-way chips, then drill evidence sentences in Blind. The planting graph will not remind you — the syllabus will.