Calories on a food label come from a surprisingly simple system: each gram of protein counts as 4 calories, each gram of fat as 9, and each gram of carbohydrate as 4. These values, known as the Atwater factors, were established in the late 1800s and remain the basis for nearly every nutrition label worldwide. But behind those round numbers is a layered process involving laboratory combustion, digestion estimates, and some known compromises in accuracy.
What a Calorie Actually Measures
A calorie is a unit of energy, defined as the amount needed to raise the temperature of one gram of water by one degree Celsius. That’s an extremely small amount of energy, so the “calories” listed on food packaging are actually kilocalories (kcal), each equal to 1,000 of those tiny laboratory calories. When a label says a banana has 105 calories, it means 105 kilocalories. Outside the United States, energy is often listed in kilojoules instead, with 1 kcal equaling 4.184 kilojoules.
Burning Food in a Lab
The original method for measuring food energy is bomb calorimetry. A small food sample, typically between 0.3 and 0.75 grams, is sealed inside a heavy steel chamber, surrounded by water, and ignited with an electrical spark in a pure oxygen environment. The food burns completely, and the resulting rise in water temperature reveals the total chemical energy stored in that sample.
Getting an accurate reading requires careful preparation. The food is first blended or minced to distribute nutrients evenly, then dried (either freeze-dried or oven-dried) to remove water, which doesn’t contribute energy but interferes with clean combustion. The dried sample is ground into a fine powder and pressed into a small pellet so it burns evenly rather than flaring up. Labs typically run each sample at least twice, and if results differ by more than a small margin, they run it a third time. The calorimeter itself is calibrated daily using a substance with a known energy value, usually certified benzoic acid.
This method captures “gross energy,” the total heat a food could theoretically release. But your body isn’t a furnace. You don’t absorb every last molecule of what you eat, which is why the calorie counts on labels use a different approach.
The Atwater System Behind Food Labels
In the 1890s, chemist Wilbur Atwater conducted hundreds of experiments measuring how much energy humans actually extract from food. He burned foods in a calorimeter, then accounted for losses through digestion and waste. The result was a set of average conversion factors: 4 calories per gram of protein, 9 calories per gram of fat, and 4 calories per gram of carbohydrate. Alcohol, when listed, adds 7 calories per gram.
These are the numbers food manufacturers use today. To calculate the calories in a packaged food, a lab analyzes the product to determine how many grams of protein, fat, and carbohydrate it contains. Those gram amounts are multiplied by their respective Atwater factors and added together. A food with 10 grams of protein, 5 grams of fat, and 30 grams of carbohydrate would be calculated as (10 × 4) + (5 × 9) + (30 × 4) = 205 calories.
This system is practical and consistent, but it treats all foods of the same macronutrient composition as equal. A gram of protein from chicken is assigned the same 4 calories as a gram of protein from lentils, even though your body may handle them differently.
Why Label Calories Can Be Inaccurate
The Atwater factors are averages, and they don’t account for how food structure affects digestion. Research on almonds illustrates this clearly. In a controlled crossover study with 18 participants, researchers measured the actual energy people absorbed from different forms of almonds: whole raw, whole roasted, chopped, and as almond butter. The Atwater system overpredicted the calories absorbed from whole and chopped almonds. Whole raw almonds delivered fewer usable calories than whole roasted almonds, likely because raw almonds are harder and break into larger pieces during chewing, trapping lipids inside intact cell walls where digestive enzymes can’t easily reach them.
This pattern extends beyond nuts. Cooking, blending, and refining foods generally makes their calories more available to your body. A raw carrot and a pureed, cooked carrot may list identical calories on a label, but you’ll extract more energy from the cooked version. The calorie system doesn’t capture these differences.
Your Body’s Cost of Digestion
Not every calorie you swallow ends up fueling your cells. Your body spends energy breaking food down, absorbing nutrients, and processing them. This is called the thermic effect of food, and it varies dramatically by macronutrient. Protein costs the most to digest, using 20 to 30% of its calorie content just for processing. Carbohydrates require 5 to 10%, and fat is the most efficient at just 0 to 3%.
This means 200 calories of grilled chicken and 200 calories of butter don’t leave you with the same net energy. After digestion costs, you retain significantly fewer usable calories from the chicken. The Atwater system doesn’t subtract these costs from the label, so the number you see always overstates what your body actually keeps from protein-rich foods relative to fatty ones.
How Calorie Burn Is Estimated
Calories don’t just describe food. They also describe how much energy your body uses. The gold standard for measuring this is indirect calorimetry, where a machine captures your breath and analyzes how much oxygen you consume and how much carbon dioxide you exhale. These gas volumes get plugged into an equation that converts them to calories burned per day. This is how hospitals and research labs measure metabolic rate with precision.
Most people don’t have access to that equipment, so calorie burn is typically estimated with prediction formulas. The most widely validated is the Mifflin-St Jeor equation, which uses your weight (in kilograms), height (in centimeters), age, and sex to estimate your resting metabolic rate. For women, the formula is (9.99 × weight) + (6.25 × height) minus (4.92 × age) minus 161. For men, it’s the same but you add 5 instead of subtracting 161. In validation studies, this equation predicted resting metabolic rate within 10% of the measured value for about 71% of participants. That’s the best accuracy among common formulas, but it still means roughly one in three people get an estimate that’s off by more than 10%.
How Fitness Trackers Estimate Calories
Wearable devices take a different approach, combining real-time sensor data with user profile information. A typical fitness tracker uses a three-dimensional accelerometer to detect movement and an optical heart rate sensor to monitor exertion. Your profile data (weight, height, sex, age) provides the baseline metabolic estimate, and the sensor data adjusts it throughout the day. When your heart rate rises and the accelerometer detects sustained movement, the device’s algorithm increases the calorie estimate accordingly.
The challenge is that calorie expenditure during exercise is influenced by factors these devices can’t easily measure, including body composition, movement efficiency, and even ethnicity. Each manufacturer uses its own proprietary algorithm, and accuracy varies across brands and activity types. These trackers are generally better at detecting relative changes (you burned more today than yesterday) than at reporting precise absolute numbers.
What the Numbers Are Worth
The calorie counts on food labels and fitness apps are estimates built on useful but imperfect systems. The Atwater factors give you a reliable way to compare foods, even if they don’t perfectly predict what your body absorbs. The form your food takes (whole versus ground, raw versus cooked) affects the real number. The macronutrient mix matters too, since protein-heavy meals cost more energy to digest than fat-heavy ones. Treat calorie counts as a solid approximation rather than an exact measurement, and they remain one of the most practical tools available for managing what you eat and how you move.