How to Track Calories Burned: Complete Guide to Accurate Monitoring
- 7 Minute Read
Activity trackers nail your heart rate but miss calories burned by 27% to 93%, a Stanford study found. Here is how to track calories burned with formulas, devices, and a method that corrects itself over time.
How to track calories burned comes down to three steps: estimate your resting and everyday calories with a formula (at the Sedentary level if you will log workouts), add workout calories from a wearable, an app, or MET values, and check the result against your weight trend over two to four weeks.
The check is what catches a bad estimate. On GLP-1 medication, where intake is small, an estimate 300 calories too high is a larger share of the budget and wipes out most of a 500-calorie deficit.
Calories burned means total daily energy expenditure, or TDEE: the calories your body uses in 24 hours. Basal metabolic rate (BMR) takes roughly 60% to 70% of it for most adults, digestion about 10%, and physical activity the rest, the part that varies most from person to person.
Typical shares of daily energy expenditure; activity is the part you can change.
BMR is what you would burn lying still all day. It scales with body size and lean mass, which is why every formula asks for weight, height, age, and sex. Resting metabolic rate (RMR) is the same idea measured under looser conditions and runs slightly higher; apps treat the two as one. No wearable measures it; a watch estimates BMR from your profile, as a formula does.
Digestion costs about 10% of the calories eaten on a mixed diet: protein 20% to 30% of its own calories, carbohydrate 5% to 10%, fat 0% to 3%. The activity multipliers below already include it, so do not add a separate line.
Physical activity has two halves: planned workouts, and non-exercise activity thermogenesis (NEAT), the walking, standing, carrying, and fidgeting that fill the rest of the day. NEAT is why two people of the same size with the same gym schedule burn different totals.
A formula estimates your caloric needs from weight, height, age, and sex. Treat the result as a starting estimate; the weight-trend check at the end of this guide corrects it.
Mifflin-St Jeor, published in 1990, works in kilograms and centimeters:
Men: 10 x weight (kg) + 6.25 x height (cm) - 5 x age (years) + 5
Women: 10 x weight (kg) + 6.25 x height (cm) - 5 x age (years) - 161
In pounds and inches: 4.536 x weight (lb) + 15.88 x height (in) - 5 x age, then + 5 for men or - 161 for women
Example: a 30-year-old woman, 143 lb (65 kg), 5 ft 5 in (165 cm).
10 x 65 = 650
6.25 x 165 = 1,031.25
5 x 30 = 150
650 + 1,031.25 - 150 - 161 = 1,370 calories per day at rest
Revised Harris-Benedict (1984), for women: 447.593 + 9.247 x weight (kg) + 3.098 x height (cm) - 4.330 x age. It puts the same woman at about 1,430 calories, 60 above Mifflin-St Jeor.
Katch-McArdle uses lean body mass alone: 370 + 21.6 x lean mass (kg); at 28% body fat her lean mass is 103 lb (46.8 kg), about 1,381 calories. It needs a body-fat measurement, which smart scales estimate less accurately than calipers or a DEXA scan.
Multiply the resting number by an activity factor. The standard factors, applied to the example woman’s 1,370-calorie BMR:
| Activity level | Factor | Covers | TDEE, example woman |
|---|---|---|---|
| Sedentary | 1.2 | Little or no exercise | 1,644 |
| Lightly active | 1.375 | Exercise 1 to 3 days a week | 1,884 |
| Moderately active | 1.55 | Exercise 3 to 5 days a week | 2,124 |
| Very active | 1.725 | Hard exercise 6 to 7 days a week | 2,364 |
| Extra active | 1.9 | Physical job plus hard training | 2,603 |
If you log workouts or sync a tracker, choose Sedentary; the higher factors already contain the exercise.
The best way to track calories burned is a combination: a formula for the baseline, one source for workout calories, and your weight trend as the check. No single tool does the job alone.
A wristband reads movement and heart rate, then runs both through the maker’s formula. A 2020 review of 158 publications across nine brands (Fuller et al., JMIR mHealth and uHealth) found no brand accurate for energy expenditure, though steps and heart rate were accurate in lab settings. The best fitness tracker for calories burned is the one you wear daily and sync to one place; a consistent error can be corrected, an occasional device cannot.
Heart rate tracks oxygen use during steady cardio, so a strap or a watch estimates a run or a ride reasonably. Heat, caffeine, stress, and the start-stop pattern of intervals and strength training raise the pulse without matching muscle work. Heart rate is measured directly; calories are inferred. Heart rate variability is a good reason to wear the strap; a precise calorie count is not.
I run MyNetDiary, so read this part with that in mind. MyNetDiary imports workout calories and steps from Apple Health, Health Connect, Google Health (the link for Fitbit and Pixel trackers), Garmin, Withings, Oura, and Whoop into the day’s Calorie Budget; Apple Health on iPhone and Health Connect on Android sync free, the direct device links are Premium. On iPhone it reads 14 Apple Health data types, workout calories included, and writes 51 back. For trackers that report a daily total (Fitbit, Garmin) the app subtracts your maintenance calories from that total and counts only the extra. Without a device, the exercise log estimates calories by activity type, duration, and your weight, and you choose the formula behind your caloric needs (the calorie-planning article lists the options).
Advanced Autopilot does the trend check for you. An adaptive algorithm: instead of relying only on a profile formula, it estimates your actual energy expenditure (TDEE) from your logged food, logged exercise, and Trending Weight, then updates calorie and macro targets as that estimate changes. Your weight forecast updates with every entry. Available in Premium. MacroFactor introduced this kind of adaptive algorithm before MyNetDiary added Advanced Autopilot. Pair it with a staff-verified food database built on USDA and NCC research-grade sources, because the food side is the bigger error.
A logged walk adds its calories to the day’s Calorie Budget when Add Exercise to Calorie Budget is on.
No device? METs are the simplest way to track burned calories by hand. A MET is a multiple of resting metabolism, from the 2024 Adult Compendium of Physical Activities: sitting is 1, brisk walking at 3.5 to 3.9 mph is 4.8, running at 6 mph (a 10-minute mile) is 9.3. Calories per minute = MET x 3.5 x body weight (kg) / 200; divide pounds by 2.2 for kilograms. For a 143 lb (65 kg) person:
Those are gross figures. Subtract 1 MET first for the exercise-only number: the walk drops to about 130 calories, the run to about 283, and that smaller number is the one to add to a food budget.
Not very. In Stanford’s 2017 test of seven wristbands, none estimated calories burned within 20% of the lab measurement; the most accurate was off by 27% on average, the least accurate by 93%. Six of the seven measured heart rate within 5%.
Stanford’s best device averaged 27% off, and the others ran worse. Three reasons. The algorithm assumes population averages, and the Stanford authors’ own explanation was that energy expenditure varies with fitness level, height, and weight. The wrist is a poor place to measure whole-body work: cycling, carrying, an incline, and a set of squats barely move it. And a watch’s daily total already includes an estimate of your resting burn, so adding it to an app budget counts that burn twice.
The Stanford study (Journal of Personalized Medicine, 2017; summary from Stanford Medicine) had a diverse group of 60 volunteers walk, run, and cycle wearing seven devices, Apple Watch and Fitbit Surge among them, checked against a medical-grade electrocardiograph and breath analysis. None measured energy expenditure well, and the authors set 10% as the error a lay user could live with. The 2020 Fuller review reached the same conclusion across 158 publications and nine brands.
How to accurately track calories burned, in practice:
Track trends. The formula, the device, the MET table, and the food label are all estimates; the scale integrates them, and over a few weeks your weight change against your average intake is the number closest to what actually happened.
Reverse the math by hand: maintenance calories equal average daily intake plus pounds lost x 3,500, divided by days. Eat 1,800 a day for 28 days and lose 2 lb: 1,800 + 7,000 / 28 = about 2,050. It is a rough check, not a measurement; water weight and logging errors distort it over short periods. Advanced Autopilot does this arithmetic from your logged food, exercise, and Trending Weight and adjusts targets as the estimate changes.
In the 2026 study we published in the Journal of Medical Internet Research (plain-language summary), among 5,132 MyNetDiary users who signed up in May 2025 with a weight-loss goal, logged food on 21 or more days, and weighed in monthly, those logging food six to seven days a week lost 11.33% of body weight on average over six months, against 6.84% for fewer than two days a week. An association, not proof of cause, in a group selected for regular logging; physical activity and medication use were not measured. The study measured logging frequency, not calorie-burn precision; the habit it links to results is logging consistently.
Estimate resting calories with Mifflin-St Jeor, multiply by an activity factor that leaves out the workouts you will log (Sedentary for most people), add workout calories from one source (a wearable, a heart-rate monitor, or MET values), and log it with your food. After two to four weeks, adjust the estimate if your weight trend disagrees.
A combination: a formula for the baseline, one device or MET source for workouts, and your weight trend as the check. No consumer device measures calories burned directly; in Stanford’s 2017 test none of seven wristbands came within 20%.
Use MET values: calories per minute = MET x 3.5 x body weight in kilograms / 200. A brisk walk is 4.8 METs, so a 143 lb (65 kg) person burns about 164 calories in 30 minutes; subtract 1 MET first to count only the calories above resting.
In a lab, indirect calorimetry, which measures oxygen consumed and carbon dioxide produced; no wristband matched it in Stanford’s 2017 test, where errors ran from 27% to 93%. For everyday use, the number that holds up is a formula adjusted against your own intake and weight trend over several weeks.
Trackers measure steps and heart rate well and calories poorly; a 2020 review of 158 studies found no brand accurate for energy expenditure. Apps put food and exercise in one balance, and in our 2026 peer-reviewed study of 5,132 engaged MyNetDiary users, a group selected for regular logging, those logging food six to seven days a week lost more over six months than those logging under two days.
Not all of them, and first make sure your plan does not already include that exercise. Device estimates run high or low (none in Stanford’s 2017 test came within 20%), and device and MET totals usually include your resting burn. For a weight-loss goal the cheaper mistake is eating back less: add net calories, eat back a share, and let your weight trend decide.