Our Quảng Trị Liberica cups at 84.62. That number is doing real work, and almost nobody knows how to read it.
Here is how the scale is actually constructed, and what to make of a score when you see one.
The 100-point scale is really a 50-point scale
The SCA cupping form has ten attributes. Each is scored from 6.00 to 10.00 in quarter-point steps. Ten attributes at a maximum of 10 gives 100.
But the floor is 6, not 0. A coffee that scored the minimum on every attribute would score 60, and nothing that bad ever gets formally cupped. In practice the entire meaningful range runs from about 75 to about 92, and specialty lives between 80 and 90.
So the working resolution is much narrower than "out of 100" suggests. A point is a lot. The difference between 84 and 86 is substantial.
The ten attributes
Fragrance/Aroma — the smell of the dry grounds, then of the wet crust after water is added. Two assessments, one score.
Flavour — the main impression across the palate. Broadest attribute and often the most heavily weighted in a grader's thinking.
Aftertaste — how long the positive flavour persists after swallowing, and whether what remains is pleasant. Coffees that fall apart on the finish get caught here.
Acidity — assessed for quality, not quantity. A grader is not rewarding more acidity; they are asking whether the acidity present is bright and structural or sour and flat. This is the attribute most often misread by consumers.
Body — tactile weight and texture. Again quality, not quantity: a light body can score well if it is silky and appropriate.
Balance — whether flavour, aftertaste, acidity and body work together or fight. A coffee can score well on every individual attribute and lose points here.
Uniformity — five cups are brewed from the same sample. Do they taste the same? Scored 2 points per consistent cup, 10 maximum.
Clean Cup — absence of interfering negative flavour, again across five cups, 2 points each.
Sweetness — perceived sweetness, five cups, 2 points each.
Overall — the grader's holistic judgement. Explicitly subjective, and where a technically sound but boring coffee gets marked down and an unusual one gets rewarded.
Note that Uniformity, Clean Cup and Sweetness are effectively pass/fail per cup. A flawless sample takes 30 points from those three almost automatically — which is another reason the practical range is compressed.
Defects are subtracted
After scoring, defects come off. A taint — a noticeable off-flavour that does not ruin the cup — costs 2 points per affected cup. A fault — an overwhelming off-flavour — costs 4 per cup.
One faulty cup out of five costs 4 points, which moves a coffee from excellent to merely good. This is why processing consistency matters as much as peak quality.
Why the decimals
84.62 looks like false precision, and in a sense it is — no palate resolves hundredths of a point.
It is usually an average. A sample is typically cupped by a panel, or the same grader cups multiple sample sets, and the reported figure is the mean. Average five graders' scores of 84.5, 84.75, 84.5, 85.0 and 84.25 and you get a number with decimals attached.
Treat the whole number as the signal and the decimals as arithmetic.
What the bands mean in practice
| Score | Band | What it means |
|---|---|---|
| 90+ | Outstanding | Rare. Competition lots, auction prices. |
| 85–89.99 | Excellent | Top of the specialty market. Distinctive and well-executed. |
| 80–84.99 | Very good | Specialty grade. The bulk of good specialty coffee. |
| 75–79.99 | Below specialty | Sound coffee, commodity grade. |
| <75 | Commodity | Blending and instant. |
So 84.62 sits near the top of the "very good" band — solid specialty, comfortably above the threshold, not claiming to be a competition lot. That honesty is the useful part.
The Fine Robusta protocol
This matters for Vietnam specifically.
Until 2010 there was no formal way to score Robusta. Graded on the Arabica form, Robusta lost automatically on acidity, because it does not have Arabica's acidity and was never going to.
The SCA's Fine Robusta protocol, developed with the Coffee Quality Institute, gave the species its own form. It replaces some attributes — Salt/Acid and Bitter/Sweet aspects appear in place of the straight Arabica acidity assessment — and redefines what a defect is in a Robusta context.
The consequence is that a Fine Robusta scoring 84 is not claiming to taste like an Arabica scoring 84. It is claiming to be excellent at being Robusta. Comparing the numbers across protocols is a category error, and a common one.
Liberica is typically assessed under the Fine Robusta protocol too, for the same reason: the Arabica form does not describe what it is doing.
What a score does not tell you
Freshness. A score is a snapshot from one sample at one moment. A coffee that cupped 86 in March and has been sitting in a warehouse since is not an 86 now.
Roast quality. Cupping uses a standardised light roast specifically to assess the green coffee. The roast in the bag you buy is a different decision by a different person, and it can waste an excellent green entirely.
Whether you will like it. The scale measures quality against a professional standard, not preference. An 87 washed Ethiopian and an 84 honey Robusta are not competing for the same moment, and the 84 may well be the one you want.
Whether the number is real. Anyone can print a number on a bag. A score means something when it names a protocol and, ideally, a grader or a lab. "SCA 86" with no provenance is a marketing claim.
Using it
Read the score as one input, not a verdict. A score plus a harvest date plus a named origin plus a described process is a coffee whose producer is telling you things that can be checked. That combination is worth much more than any of the parts, and it is the practical difference between specialty and commodity.

