Coffee Roast Weight Loss Percentage: What 70 Logged Home Roasts Show

A light filter roast typically loses about 12% to 16% of its green weight. In 70 logged home roasts on a Kaffelogic Nano, almost every roast lost between 11.67% and 16.7%. Roasts that ended around 213 °C lost about 12% to 14.5%; roasts that ended around 223 °C lost about 13.6% to 16.7%. The bean decides where in that range you land almost as much as the roast does, so compare a roast's weight loss with other roasts of the same coffee, not with a chart.

The rest of this page shows how to calculate it, the numbers behind those ranges, and what weight loss can and cannot tell you.

How to calculate roast weight loss

Weigh the green beans before the roast and the roasted beans once they have cooled:

weight loss % = (green weight − roasted weight) ÷ green weight × 100

120 g green → 102.6 g roasted
(120 − 102.6) ÷ 120 × 100 = 17.4 ÷ 120 × 100 = 14.5%

Most of what leaves is water, plus some chaff and the gases and volatiles driven off as the beans brown. Use the same scale both times, weigh to 0.1 g if you can, and weigh after the beans have cooled, because chaff and moisture keep coming off for a few minutes.

Where these numbers come from

The roasts in this article are the ones behind our development time ratio study: light roasts for filter of coffees from Kenya, Ethiopia, Colombia, Honduras and a Brazil decaf, each logged with its profile, drop temperature, weight lost and a 0–5 cup score. 70 of them have a weight loss recorded. The repeat-roast figures further down come from a separate export of the recipes that were roasted several times at the same profile, level and load.

They are light roasts only. The hottest drop in the set is about 225 °C, so this data says nothing about medium-dark or dark roasts, which lose more.

Weight loss by bean and drop temperature

Two things stand out.

Within one bean, a hotter drop loses more. The Yirgacheffe lost 12.33–13.17% on the profile that ended near 214 °C and 13.6% on the one that ended near 222 °C. The Colombia lost 14.17–14.5% at about 213.8 °C and 15.3–16% at about 223 °C. Its hottest roast in the set, R-020 at 223.55 °C, lost 16.7%.

Across beans, the same temperature gives different losses. At almost the same drop temperature, the Colombia R-051 (213.91 °C) lost 14.17% while the Kenya R-044 (213.97 °C) lost 11.92%: more than two points apart. At the hot end, the Brazil decaf R-052 ended at 223.5 °C and lost 14%, while the Colombia R-020 ended at 223.55 °C and lost 16.7%. Moisture content, density and processing all differ between lots, and the scale sees all of it.

That is why a generic chart of "light = X%, medium = Y%" is a weak guide for a home roaster. A 14% loss is a hot roast for the Kenya above and a cool one for the Colombia.

How much weight loss varies between repeat roasts

On a fixed recipe (same bean, profile, level and batch size), weight loss repeats closely:

The widest spread is 1.25 points (13% to 14.25%, the Cauca decaf). Most recipes stay within half a point. So when a repeat of a known recipe comes out two or more points away from its usual loss, something changed: the batch size, the green coffee (a new bag or an older one that has dried out), the room temperature, or the weighing.

This is the most useful job weight loss does at home. It is a cheap check that a roast repeated, and it catches mistakes in the log.

Weight loss does not predict the cup

The eleven Kenya roasts on F-WSH 1500-2200 all lost between 11.67% and 12.67%, a one-point window. Their cup scores ran from 1.5 to 5. R-087 lost 11.67% and scored 5; R-014 lost 12.67% and scored 1.5. The same holds in the other direction: two Colombia roasts at 16% and 16.7% scored 4.5 and 1.5.

Weight loss tracks how far the roast went, not whether that was the right place to stop for this coffee. To decide that, taste the roast after it has rested. Our rest-day data shows why an early cup misleads, and first crack is the other landmark worth logging next to the weight.

When the number looks wrong

Two roasts in the set read far outside their beans: R-101 at 9% and R-099 at 26.42%. Neither has a drop temperature logged, and both sit well away from every other roast of the same coffee. A reading like that is almost always the scale, a missed tare, or a typo, not a real roast. Before you change a recipe because of one odd weight loss:

1. Re-weigh the roasted coffee if you still have it.
2. Check the green weight you entered against the load you actually used.
3. Compare with the last roast of the same recipe. If that one was normal and the curve looks the same, trust the curve.

How to use weight loss in your own roast log

• Record green and roasted weight on every roast, and let the log compute the percentage. Typing the percentage by hand is where the 26% entries come from.
• Compare within a bean. Build up three or four roasts of a coffee and you have its own baseline; judge new roasts against that.
• Pair it with drop temperature. Weight loss plus the end temperature tells you whether a change in the cup came from a change in the roast.
• Let taste decide. Weight loss confirms a repeat; it never tells you a roast is good.

BrewedLate's roast tracker stores green and roasted weight per roast and shows the roast loss next to the profile, level and tasting, and the free Kaffelogic log viewer draws a .klog's curve in the browser. For more on getting started, see the home coffee roasting hub and the beginner's guide.

FAQ

What is a normal weight loss percentage when roasting coffee?

For light filter roasts, the roasts in our logs lost between 11.67% and 16.7% of their green weight. Roasts that ended near 213 °C lost about 12% to 14.5%, and roasts that ended near 223 °C lost about 13.6% to 16.7%, depending on the bean.

How do you calculate coffee roast weight loss?

Weigh the green coffee before the roast and the roasted coffee after it cools. Weight loss % = (green weight − roasted weight) ÷ green weight × 100. A 120 g batch that comes out at 102.6 g lost 17.4 g, which is 14.5%.

Does a higher weight loss mean a darker roast?

For one bean, yes: within the same coffee, the roasts that ended hotter lost more weight. Across beans it is not reliable. In our logs a Colombia and a Brazil decaf both ended at about 223.5 °C, and one lost 16.7% while the other lost 14%.

Does weight loss tell me if the roast will taste good?

No. Eleven roasts of one Kenya recipe all lost between 11.67% and 12.67%, and their cup scores ran from 1.5 to 5 out of 5. Use weight loss to check that a roast repeated, and taste to judge it.

What does a weight loss far outside the usual range mean?

Usually a weighing or entry mistake rather than a real roast. Two of the roasts in our logs read 9% and 26.42%, far from the rest of their beans, and both lack a logged drop temperature. Check the scale and the entry before you change the recipe.