Light, Medium, Dark

Phenomenon

You buy a bag labelled "medium roast" from each of three shops and pour them onto a sheet of white paper at home.

One is light chestnut with a dry surface. One is chocolate brown. The third is darker still, with a faint sheen of oil. All three say "medium".

At a fourth shop, the bag has no "light, medium, dark" at all. Instead it prints a number, or a row of small dots from light to dark with one of them filled in.

So which color is "medium"? If three words cannot pin it down, how should roast degree be described?

Variables

These are the readings people commonly use to describe how far a bag was roasted:

All of these readings change continuously. They move together along the same timeline, just at different rates.

Experiment

Thought experiment: put one batch of green coffee in a roaster. Starting at first crack, take out a small handful every thirty seconds — ten handfuls in all — and line them up in order on white paper.

Write down your prediction before reading on

Will the colors of the ten handfuls fall into three clear groups, or shade gradually from one to the next? If ten people each drew two lines on this row to divide "light, medium, dark", would they draw them in the same places?

Your prediction

Save prediction

Stored only in this browser on this device. Never uploaded.

A second thought experiment: two roasters both sell a "medium roast". One mainly roasts for pour-over; the other mainly roasts espresso blends.

Write down your prediction before reading on

Will their "medium" be the same color? Which one do you expect to be darker?

Your prediction

Save prediction

Stored only in this browser on this device. Never uploaded.

Observation

Lined up, the ten handfuls form a gradient with no steps: each is a little darker than the one before, and there is no point where the color suddenly jumps. As a roast progresses, weight loss, volume expansion, porosity and color all change continuously, and the darker the roast, the larger the change Confidence: [Strong Evidence]. "Light, medium, dark" are lines we draw on that gradient afterwards.

Ten people will most likely draw their lines in different places. "Light, medium, dark", and names such as "City" or "French" roast, have no single standard that everyone follows Confidence: [Common Practice]; usage varies between countries, between roasters, and even between product lines at the same shop.

The second experiment is similar. Many roasters roast beans meant for espresso a little darker, and beans meant for pour-over a little lighter Confidence: [Common Practice] (Chapter 31), so the two shops' "medium" are probably not the same color.

So the more careful practice in the industry is to measure. Roasters often measure the color of beans or grounds with a dedicated roast color meter that gives a number; on the instrument used in one study, a light roast read about 131 and a dark roast about 78 — the smaller the number, the darker Confidence: [Common Practice]. But different brands and models of color meter use different scales, so numbers are only comparable on the same kind of instrument, and whole beans read differently from grounds Confidence: [Rule of Thumb].

Model

One spectrum, a few landmarks

Think of roast degree as a continuous ruler, not three boxes. There are a few landmarks on the ruler that everyone recognizes:

• Start of first crack: the beans swell and give a sharp popping sound (Chapter 14);
• Around the end of first crack: many light roasts stop in this region;
• Start of second crack: a finer, quieter crackling; many dark roasts stop around or after second crack Confidence: [Common Practice];
• Oil on the surface: oil inside the bean is pushed to the surface, which usually means a very dark roast, or beans that have been sitting for a while Confidence: [Rule of Thumb].

"Light, medium, dark" are three ranges marked on this ruler, and where the boundaries fall depends on who is drawing them. To describe a bag, it is better to say "which landmark it is near" than "which box it belongs to".

What changes along the spectrum

Moving along the ruler from light to dark, the changes fall into two groups:

The bean's physical properties. The darker the roast, the more weight the bean loses, the larger its volume, the lower its density and the more pores it has Confidence: [Strong Evidence]. That is why a scoop of dark roast weighs less (Chapter 22), and why dark beans are more brittle, easier to grind and usually easier to extract (Chapter 14). The darker the roast, the more carbon dioxide is trapped in the bean and the faster it degasses Confidence: [Strong Evidence] (Chapter 17).

The bean's chemistry. Chlorogenic acids keep breaking down as the roast goes on, so the darker the roast, the less remains, while caffeine stays largely stable Confidence: [Strong Evidence]. Roasty and bitter flavors from the roast itself increase, and the bean's original fruit acidity and floral notes decrease. The details of flavor are in Chapter 16.

There is no "correct" box on the spectrum

Every position on the spectrum has people who love it. Whether the brightness and fruit acidity of a light roast or the weight and roasty bitterness of a dark roast tastes better is a personal preference Confidence: [Preference].

But "preference" does not mean "anything goes". Beans roasted too lightly, without enough development, often taste of grass or raw peanuts and are hard to extract; beans roasted too far turn burnt, smoky and bitter, and the character of the coffee itself almost disappears Confidence: [Rule of Thumb]. These two extremes are "roasted badly", which is a different thing from "light roast" or "dark roast".

Color is just one reading

As Chapter 14 put it: the endpoint is not the path. Two bags of exactly the same color can taste different because they were heated at different rates. So color is the most useful reading on the spectrum, but not the whole story: it tells you where the beans stopped, not how they got there.

Transfer

Given an unfamiliar bag, you can read where it sits on the spectrum and decide where to start brewing:

These adjustments are starting points, not rules Confidence: [Rule of Thumb]. Change one variable at a time (Chapter 24). If you scoop rather than weigh, remember that a scoop of dark roast is lighter (Chapter 22).

Misconceptions

"Dark roast has more caffeine." Caffeine stays largely stable through roasting Confidence: [Strong Evidence], so by weight light and dark roasts differ little. What really changes is how you measure: dark beans are less dense, so a scoop of dark roast weighs less and holds less coffee. Dark roast tastes "stronger" because of roasty and bitter flavors, not caffeine (Chapter 14).

"Light roast is under-roasted; dark roast is burnt." Light and dark are positions on the spectrum; under-development and burning are roasting badly. A good light roast can be clean, sweet and bright; a good dark roast can be heavy without tasting burnt.

"Medium means the same everywhere." There is no single standard. "Medium" from different shops or countries can differ a lot. To compare, look at the color, at roast color numbers (from the same kind of instrument), or simply ask the roaster.

"Oily beans are good beans." Oil usually means a dark roast, or beans that have sat for a while. On its own it says nothing about quality.

"Same color, exactly the same roast." Color is the most useful reading, but the same color can be reached by different timelines and taste different (Chapter 14).

Real Lab

There is no registered lab card for this chapter. Get three bags at different roast levels (one light, one medium, one dark — from the same shop if possible), a sheet of white paper, a kitchen scale and a measuring spoon, and use your phone camera if you can.

1. Sort. Pour a small pile from each bag onto the white paper and line them up in daylight. Then pick out the lightest and the darkest bean from each bag: are the beans within one bag all the same color?
2. Find the boundaries. If you have more bags (from friends, or shop samples), slot them all into the row. Try drawing two lines to divide "light, medium, dark", then ask someone else to draw their own lines independently. Did you draw them in the same places?
3. Weigh. Take one level spoonful of each with the same spoon and weigh it. Which spoonful is lightest?
4. Grind. Grind a little of each and compare the color of the grounds, and how grinding feels and sounds. Are the grounds the same color as the whole beans?

Before you start, write down your predictions: in what order will the three spoonfuls weigh? How far apart will your lines and the other person's be?

Check Your Model

Predict. The shop you usually buy from moves one coffee from "light" to "medium-light". The roast color number on the bag goes from 95 to 88 (same instrument; smaller means darker). Before looking at any result, write down: brewed with the same recipe, how might the cup change?

Diagnose. A friend buys a bag labelled "dark roast" and finds it is only medium brown, with no oil on the surface. He feels cheated. What do you think?

Explain. In your own words, explain why roast degree is a spectrum rather than three boxes.