How Food Calories Are Actually Calculated: the Atwater Method
In short
The calorie count on a label is almost never measured food-by-food in a bomb calorimeter. It's calculated from carbohydrate, protein and fat content, multiplied by fixed factors (4, 4 and 9 kcal/g) set by chemist William Atwater in the late 19th century, still used today by the FAO.
A calorie count on a label feels like a precise, near-scientific measurement, down to the gram. In the vast majority of cases, it isn't a direct measurement: it's a calculation, made from the food's macronutrient composition, using a method that's over 130 years old.
Understanding this method — known as the "Atwater method" — helps you see what a calorie figure actually represents, and why two foods with "the same number of calories" aren't always measured the same way.
A calorie isn't measured bite by bite
Directly measuring a food's energy content is done with a bomb calorimeter: the sample is burned inside a sealed chamber, and the temperature rise of the surrounding water is measured to work out the total energy released by combustion. The problem: this measures the food's raw chemical energy, not the energy the human body actually extracts from it, since some of it is never digested or absorbed. To correct for this gap, American chemist William Olin Atwater developed, in the late 19th century, a system of conversion factors that accounts for real digestibility.
The Atwater method: the 4-4-9 system
The principle rests on three average factors, one per macronutrient, expressed in kcal per gram:
- Carbohydrates: 4 kcal/g
- Protein: 4 kcal/g
- Fat: 9 kcal/g
These values are documented in the FAO technical report on food energy, methods of analysis and conversion factors, which uses the Atwater coefficients as an international reference still applied today. To get a food's calorie total, each macronutrient (in grams per 100g) is multiplied by its factor, then the three results are added up. Alcohol follows the same logic, with its own factor of 7 kcal/g.
Ready to try Radish?Private calorie counter, Nutri-Score, no ads. Free.Get Radish →Why 4-4-9 is an average, not an exact measurement
These three numbers are averages calculated across the whole range of foods, not universal physical constants. According to FAO technical documentation and food-database compilation work, the energy actually released per gram of fat can vary significantly from one food to another depending on the type of fatty acids present; the same variability exists, to a lesser degree, for carbohydrates and protein, depending on their own digestibility (non-digestible fibre, resistant starch, partially absorbed protein). That's why some nutrition databases use food-specific factors instead of the general 4-4-9 system, particularly for foods high in fibre or in poorly digestible protein.
The special case of fibre
Dietary fibre raises a methodological problem: it isn't digested in the small intestine, but partly fermented by the gut microbiota in the colon, which releases residual energy. Calculation systems differ by country: some databases assign fibre a zero energy factor, others a reduced factor (around 2 kcal/g) to reflect this partial fermentation. This is a common source of small discrepancies between national databases on fibre-rich foods, without either method being "wrong" — they simply follow different conventions.
What this means for the label you're reading
The calorie figure on a package is therefore, in the vast majority of cases, neither measured by calorimetry on the finished product nor guaranteed to the exact kcal: it's an estimate calculated from the declared macronutrient composition, with an error margin inherent to the method itself. Regulations (both EU and US) actually tolerate labelling error margins precisely for this reason: natural batch-to-batch variability adds to the variability of the calculation method itself.
How Radish shows these values
Radish never invents a calorie value: the figures shown come from official databases — CIQUAL for France, USDA for the US — which apply the Atwater method or food-specific factors themselves, following their own published protocols. When a precise value is missing for a food, the app flags it rather than silently estimating a number.
Why this matters for interpreting a food diary
Knowing that calorie figures are calculated, not measured, changes how you should read small differences between two similar foods in a tracking app. A gap of 5-10 kcal per 100g between two databases for the same generic food (say, "cooked white rice") isn't necessarily an error: it can reflect a different water content in the sample analysed, a different assumed cooking method, or a different rounding convention on the fibre factor described above. What matters for day-to-day tracking is consistency — using the same database over time so that your own trend stays comparable — rather than chasing single-digit precision that the underlying method was never designed to guarantee in the first place.
Ready to try Radish?Private calorie counter, Nutri-Score, no ads. Free.Get Radish →FAQ
Are a food's calories directly measured?
Rarely. Most values shown are calculated from carbohydrate, protein and fat content, multiplied by the Atwater factors (4, 4 and 9 kcal/g), not measured by combustion on the finished product.
Why does fibre raise a particular issue?
Because it isn't digested in the small intestine but partly fermented in the colon. Depending on the database, it's assigned a zero or reduced energy factor, which explains small discrepancies between countries on fibre-rich foods.
Is the Atwater method wrong?
No, it's a useful, documented approximation used by nearly every official nutrition database. It carries a known error margin, mainly linked to variability in real digestibility from one food to another.
The Atwater method isn't a hidden flaw: it's the standard used by nearly every official nutrition database in the world, documented and treated as a useful approximation rather than an exact measurement. In Radish, the values shown come from official databases (CIQUAL, USDA) that apply this method to international standards — never recalculated on the fly. Try the app, free, every value shows its source.
General information, not medical advice.
🔎 Methodology & sources. Every nutrition figure in this article is cross-checked against official sources (ANSES, EFSA, WHO, CIQUAL/USDA, French Public Health agency), and health topics are reviewed before publication against our internal reference framework. See our full methodology →
Prefer the installed app? Get Radish on iPhone or Android →