Should you eat back your exercise calories?
Short answerNo
On this page
No. Your usual workouts belong inside your maintenance number, so set it from your steps and your weight trend and leave the watch’s exercise calories out. Watches overstate workouts, and your body makes up part of what you burn. On a rare big day, like a three-hour hike, eat a little more.
What we’d do
Set your maintenance once with your normal week of workouts in it, from the calories calculator or from four weeks of your own logs. Then eat to that number every day, training day or not. Turn off the setting in your food app that adds exercise calories to your budget. On an unusual day, a long hike or a race, eat back about half of what the watch says. If you’re losing faster than 1% of your weight a week or feel wiped out in your workouts, raise your daily number by 100 to 200 instead.
If you remember three things
- Your maintenance already counts your usual workouts. Adding them again counts them twice.
- Watches overstate calories. In a Stanford test of seven devices, median errors ran from 27% to 93%.
- Your body takes some back. About 28% of activity calories, in a study of 1,754 adults, plus the extra hunger.
Why it sounds right
The logic is clean. If you burned 400 extra calories on a run, you have 400 more to spend and you’re still in the same deficit. Many food apps do exactly this, adding your workout to the day’s budget the moment your watch syncs.
The logic is fine, and the inputs are wrong. Eating back a workout is like spending a check before it clears, when the check is often written for more than the bank will pay. The workout number is the check. What actually lands in your account is smaller, and sometimes much smaller.
Three reasons it backfires
It counts your workouts twice. A maintenance number from a calculator includes your activity level. One from your weight trend includes every workout you did that month. A watch’s active calories are, in Apple’s own definition, “the energy that the user has burned due to physical activity and exercise”, on top of resting energy. Add them to a maintenance that already has them, and you’ve eaten your workout twice.
The number is too big. In the 2017 Stanford test, 60 people wore seven wrist devices while their real burn was measured from their breathing. No device got calories within 20%, and errors were highest during walking. A 2025 review of 56 Apple Watch studies found calories missed the 10% mark in every group of studies, while heart rate never did.
Your body makes some of it back. In a 2021 study of 1,754 adults, only about 72% of the extra calories burned in activity showed up as extra calories burned that day. In a 24 week trial, people who did supervised exercise ate about 90 to 124 calories a day more than a group that didn’t.
What the studies found
Exercise loses less than the math says
171 adults with overweight or obesity, 24 weeks, 2019. Martin and colleagues, E-MECHANIC trial
Two groups exercised under supervision at two doses. Both lost less than their workouts predicted, by about 1.5 kg and 2.7 kg. They ate more, and their hunger rose. Their resting burn and their movement outside workouts didn’t change.
Mostly women. Food intake was worked out from doubly labeled water and body changes, not logged.
Why exercise trials disappoint
A review of exercise trials that tracked how much people actually did, 2012. Thomas and colleagues, Obesity Reviews
The small weight loss in most exercise trials came mainly from small doses of exercise and an increase in eating that came with them.
Built from trials with different designs and lengths.
The body trims elsewhere
1,754 adults, 2021. Careau and colleagues, Current Biology
About 28% of activity calories were offset by a lower resting burn. People with more bodyfat offset more.
Across people at one point, not the same people before and after.
Put it together. Say your maintenance is 2,200 with your three weekly runs in it, and your watch credits each run with 450 calories. Eat those back and you’ve added 1,350 calories a week. On a deficit of 500 a day, that cancels almost 40% of it, and you’d lose about 40% slower than planned.
Where it’s partly right
Big, unusual days are real. If you hike for three hours or run a half marathon, you burn far more than on a normal day. That’s why we’d eat back about half on those days: enough to recover, not enough to erase the week’s deficit. And people training hard for an event, many hours a week, need more food than a desk worker, which their maintenance should show.
The bigger the training load, the less each extra hour adds. In a 2016 study of 332 adults, daily burn rose with activity at the low end and levelled off at the high end. The Hadza, hunter-gatherers whose men walk about 7 miles a day, burn about what Westerners of the same size burn. Exercise is worth doing for many reasons. Eating it back isn’t one of them.
More detail: what to do if you track with a watch
- Set your food app’s daily target as a fixed number, from your measured maintenance minus your deficit, and switch off any setting that adjusts it for exercise.
- Use the watch’s step count to keep your daily movement steady. That matters more than the workout number.
- If you want to see what your watch gets wrong for you, compare its average daily total over four weeks with your measured maintenance.
8 sources
- Apple, HealthKit: activeEnergyBurneddeveloper.apple.com
- Shcherbina et al., Accuracy in wrist-worn, sensor-based measurements of heart rate and energy expenditure in a diverse cohort, Journal of Personalized Medicine, 2017pmc.ncbi.nlm.nih.gov
- Choe and Kang, Apple Watch accuracy in monitoring health metrics: a systematic review and meta-analysis, Physiological Measurement, 2025pubmed.ncbi.nlm.nih.gov
- Careau et al., Energy compensation and adiposity in humans, Current Biology, 2021pubmed.ncbi.nlm.nih.gov
- Martin et al., Effect of different doses of supervised exercise on food intake, metabolism, and non-exercise physical activity: the E-MECHANIC randomized controlled trial, American Journal of Clinical Nutrition, 2019pubmed.ncbi.nlm.nih.gov
- Thomas et al., Why do individuals not lose more weight from an exercise intervention at a defined dose? An energy balance analysis, Obesity Reviews, 2012pubmed.ncbi.nlm.nih.gov
- Pontzer et al., Constrained total energy expenditure and metabolic adaptation to physical activity in adult humans, Current Biology, 2016pubmed.ncbi.nlm.nih.gov
- Pontzer et al., Hunter-gatherer energetics and human obesity, PLoS One, 2012pmc.ncbi.nlm.nih.gov
