Sensium confusion neighborhoods 2026: the 15 most-planted grapes and who they collide with

Grape confusion pairs are usually taught as a flat list. Serious tasters also need neighborhoods: for each high-planted grape, which four catalog look-alikes sit on its dossier, which edges are mutual, and which separator decides the call. On 2026-07-09 we exported those neighborhoods for the 15 most-planted grapes in the Sensium catalog — the same `confusionPairs` graph that powers Compare, Train suggestions, and every client’s “most confused with” chips.

Headline finding: the densest planting ranks are not lonely stars. Each of the top 15 carries a four-neighbor editorial cap, and for ranks 1–15 those neighbors are almost entirely mutual — both dossiers list each other. That means the product surfaces a cluster, not a one-way tip. Study 1a ranked edges by planting weight (confusion graph 2026); this Study 1b ranks hubs by planting rank and shows the local map around each hub.

This is still catalog-graph research, not anonymized user telemetry. Population miss-rates remain Study 2 when volume clears. Until then, cite neighborhoods as editorial graph structure — reproducible, dated, and honest about what they are not.

Methodology (read this before citing the tables)

What this ranking is not: a claim that these are the grapes candidates miss most often, or that four neighbors are the only possible confusions in a flight. It is the local editorial map the product already uses when you open a dossier or Compare from that grape.

Companion: Study 1a — top 25 pair ranks. Use 1a for “which edge first?”; use 1b for “I have Cabernet / Chardonnay / Tempranillo in the glass — which four twins do I rehearse?”

Why neighborhoods matter more than another top-25 list

A ranked edge list answers priority. A neighborhood answers study design. If you only drill Cabernet vs Merlot, you still fail Cabernet vs Syrah, Cabernet vs Malbec, and Cabernet vs Cabernet Franc on the same week’s paper. The catalog already stores those four edges on the Cabernet dossier — publishing them as a hub makes the drill plan obvious: open the hub, walk the four neighbors, write one separator each.

Google’s 2026 guidance on generative AI search still rewards non-commodity pages: unique structure, auditable method, proprietary data. A recycled “grapes people confuse” listicle is commodity. A dated neighborhood export with mutual flags and a re-run command is not.

The 15 hubs (planting ranks 1–15)

Separators are the catalog’s first editorial cue for that edge (trimmed). Open Compare for the full fingerprint. M = mutual; → = one-way from the hub.

Reds — Bordeaux, Iberia, Rhône spine

Drill tip for this block: treat Tempranillo–Syrah–Grenache–Sangiovese as a clique. If you miss one edge, schedule the other three the same week — the neighborhoods overlap on purpose.

Deep dives already on site: Cabernet vs Merlot, Syrah vs Grenache, Mourvèdre vs Syrah, Carignan vs Grenache, Tempranillo vs Sangiovese, Cinsault vs Pinot.

Whites — noble + high-planted neutrals

Drill tip for this block: Airén–Trebbiano–Macabeo is the neutral-white triangle Study 1a already flagged in the top 25. Chardonnay’s neighborhood is not “only Sauvignon” — it also collides with those neutrals and with Pinot Gris. If your syllabus is noble-only, still open Airén once from this hub table.

Deep dives: bright whites trio, Sémillon vs Sauvignon, Chenin vs Chardonnay, Pinot Gris vs Grigio, Chardonnay three faces.

Three takeaways from the neighborhoods

1. Mutual is the default at the top of the planting list

For ranks 1–15, the exported neighbors are overwhelmingly mutual. That is a graph-health check: the densest commercial grapes are not given one-way “also try” tips; they are wired as reciprocal confusion. When you see a → one-way edge lower in the catalog (Nebbiolo → Pinot, Pinotage → Syrah), treat it as a different study object — exam-critical, but not planting-weighted mutual spine.

2. Hubs overlap into cliques — drill the clique, not one edge

Cabernet’s neighborhood shares Merlot and Malbec with Merlot’s hub; Tempranillo’s neighborhood shares Syrah/Grenache/Sangiovese with those hubs; Airén’s neighborhood shares Trebbiano/Macabeo with Chardonnay’s white collisions. Misses compound. After any wrong call, open the hub of the grape you named, not only the pair you meant.

3. Color-frame edges still appear inside red/white hubs

Pinot Noir ↔ Pinot Gris sits in both neighborhoods. The first separator is frame (red vs white), not a mid-palate nuance. The graph keeps the DNA/family edge so Compare does not hide it; your drill should still start with color before cleverness.

How to drill neighborhoods in four weeks

1. Week 1 — hubs #1–3 + #6–7. Cabernet, Merlot, Tempranillo, Syrah, Grenache. Four neighbors each; log separators in Train.
2. Week 2 — hubs #9, #11–13. Pinot, Sangiovese, Bobal, Carignan. Add Cinsault vs Pinot when pale perfume appears.
3. Week 3 — hubs #4–5, #8, #10, #14–15. Airén, Chardonnay, Sauvignon, Trebbiano, Macabeo, Pinot Gris. Force one neutral-white flight.
4. Week 4 — revisit misses + one exam one-way hub. Open Nebbiolo or Pinotage dossiers for one-way neighborhoods the planting list under-weights — then return to mutual spine.

Companion method: structure-first, flashcards vs drills, year-in-review pairs hub, French / Italian / Spanish place hubs when the miss is place, not cultivar.

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 beside Study 1a edges and these 1b hubs — same catalog objects, different question (“what do tasters actually confuse?”). Until that ship date, cite this page as catalog-graph neighborhoods, not population accuracy.

Frequently asked questions

How is this different from the confusion graph 2026 post?

Study 1a ranks pairs (edges) by mutual + combined planting. This Study 1b ranks hubs (the 15 most-planted grapes) and lists each hub’s local neighbors. Same graph, different cut.

Why only four neighbors per grape?

The editorial confusion graph caps dossier chips at four high-signal targets so Compare and mobile UI stay teachable. Neighborhoods inherit that cap — they are the product map, not an exhaustive ontology.

Why isn’t Nebbiolo a hub here?

Nebbiolo’s planting rank sits outside 1–15. Its neighborhood is exam-critical (often one-way into Pinot/Sangiovese/Barbera) and has deep dives — Nebbiolo vs Pinot — but this export is deliberately planting-hub first. For the curated one-way exam list (Nebbiolo, Pinotage, Carmenère, …), see one-way exam hubs 2026.

Can I cite these neighborhoods in a class or newsletter?

Yes — attribute “Sensium confusion neighborhoods, 2026-07-09” and link this URL. Do not restate them as “most-missed exam grapes” unless you are citing a future Study 2 that says so explicitly.

How do I reproduce the tables?

From a Sensium checkout: `node scripts/data/export_confusion_neighborhoods.mjs --limit 15`. Pair ranks: `node scripts/data/export_confusion_graph_ranking.mjs --limit 25`.

Open hub #1 in Compare from Cabernet Sauvignon, walk its four neighbors, then move to Merlot’s hub — overlapping edges are features, not duplicates. Neighborhoods only help if they become reps in Train and Blind.