# Should You Eat Back Exercise Calories? What the Research Shows _Only if your calorie target leaves exercise out. Why watch burns run high, how much the body offsets, and a simple two-week test to settle it._ **The short answer** **Only eat back exercise calories if your calorie target leaves exercise out.** If your activity level already includes your workouts, as most "lightly active" and higher settings do, adding the watch number counts them twice. If your target assumes no exercise, eating back part of the burn is reasonable, then check your two-week weight trend and adjust. - **20%+** - Calorie error on every wrist device tested - **28%** - Activity calories the body offsets - **1 MET** - Resting burn, already in your daily total - **2 weeks** - Time to test any change against your trend Your watch says a run burned 450 calories, and your app adds those 450 to today's budget. It feels like a reward, but if the scale has stopped moving, those extra calories are one of the first places to look. Whether to eat back exercise calories is mostly a bookkeeping question: was that burn already counted somewhere else, and how accurate is the number in the first place? Research on fitness trackers, energy use and food logging answers both. This guide shows how to tell if your target already includes exercise, why burn estimates run high and what the body does with extra activity. It ends with a simple rule and a two-week test that works with any app. ## What Eating Back Calories Means Eating back exercise calories means adding the calories a workout burned to that day's food budget. If your target is 1,700 and a ride burned 300, the app raises your budget to 2,000. Some apps call the result net calories: food eaten minus exercise logged. On paper that makes sense, since a calorie deficit is the gap between what you burn and what you eat. The trouble is that both sides of that sum are estimates, and the exercise side is often counted twice. ## Check Whether Your Target Already Counts Exercise Most calorie targets start with your resting needs, called BMR, then multiply by an activity level to get total daily needs. That multiplier is where exercise already lives. For example, "moderately active" in the [federal activity definitions](https://odphp.health.gov/our-work/food-and-nutrition/2015-2020-dietary-guidelines/advisory-report/appendix-e-3-1-a3) means activity equal to walking about 1.5 to 3 miles a day, on top of normal daily living. So if you chose an active level because you work out, your target already includes those workouts. Adding the watch's burn on top counts the same session twice. That double count can quietly erase a deficit. | How your target was set | What to do with logged exercise | | --- | --- | | Activity level includes your workouts | Don't add It's already counted | | Sedentary level, workouts logged separately | Add part Then test it against your trend | | Long training days (90+ minutes) | Eat more Fuel the session, see below | ## Why Burn Estimates Run High Even when the counting is right, the number itself is shaky. Stanford researchers tested seven wrist devices on 60 people against lab measurements. No device estimated calories burned within 20% error, though heart rate was far more accurate, according to a [2017 accuracy study](https://doi.org/10.3390/jpm7020003). A later review pooled 60 validation studies of wrist and arm monitors. Accuracy changed a lot with the type of activity, and devices that added heart rate to motion sensing made smaller errors, based on a [2020 meta-analysis](https://doi.org/10.1136/bjsports-2018-099643). So one workout reading can miss widely either way. ### Gross Calories vs Net Calories A quieter error comes from METs, the measure most calorie estimates use for how hard an activity is compared with resting. One MET is about 1 calorie per kilogram of body weight per hour, per the [2024 activity compendium](https://doi.org/10.1016/j.jshs.2023.10.010). The compendium rates general jogging at 7.5 METs. For a 154-pound person, 30 minutes works out to about 260 calories, but roughly 35 of those would have been burned sitting still and are already in your daily target. The extra cost of the jog is closer to 230. Where a 30-minute jog's calories go (154-lb person) Gross estimate, 7.5 METs**260** Minus resting calories already counted**230** After 28% average compensation**165** Illustration from the compendium's MET value and Careau et al. (2021). Results vary by person. ## The Body Gives Some of It Back Burning more doesn't add calories to your day one for one. A study of 1,754 adults found the body offsets about 28% of extra activity calories by lowering its resting burn. Only about 72% of the extra burn shows up in the day's total, according to [energy compensation research](https://doi.org/10.1016/j.cub.2021.08.016). An earlier [study of 332 adults](https://doi.org/10.1016/j.cub.2015.12.046) saw the same pattern. Total daily energy use rose with activity at lower levels but leveled off at higher ones. Exercise still works. In one 10-month trial, people told to keep their usual diet burned 400 or 600 calories per session, five days a week. They lost about 8.6 and 11.5 pounds, while a control group gained 1.1 pounds, per the [Midwest Exercise Trial](https://doi.org/10.1002/oby.20145). Still, exercise-only weight loss usually falls short of predictions, mainly because doses are small and intake creeps up, according to a [2012 energy-balance review](https://doi.org/10.1111/j.1467-789X.2012.01012.x). Food logs carry error too. In a classic study, people who said they couldn't lose weight on a low intake underreported their food by 47% and overreported their exercise by 51%. The group was small, but the [1992 NEJM study](https://doi.org/10.1056/NEJM199212313272701) shows both sides of the ledger can drift. Weighing food, as the [raw vs cooked guide](/raw-vs-cooked-weight/) explains, tightens that side. ## A Count-It-Once Rule and a Two-Week Test All of this points to one simple rule: count exercise once. Either your activity level includes it, or you log it and add it back, but never both. Then let your weight trend decide whether the setup is right. - **Pick one method.** Set your activity level to match a normal week, workouts included, and stop adding exercise calories on top. - **Hold it for two weeks.** Daily weight swings hide fat loss, so compare [weekly weight averages](/why-weight-fluctuates-daily/) rather than single mornings. - **Read the trend.** Losing about as fast as planned means keep going, while a flat trend means eat a little less. If it's falling much faster and you feel drained, eat a little more. - **Adjust in small steps.** Move your target by 100 to 200 calories, then test for two more weeks. **Long or hard training** Endurance athletes and anyone training 90 minutes or more need to fuel those sessions. Eating too little for the training load can harm health and performance, as the [IOC consensus statement](https://doi.org/10.1136/bjsports-2023-106994) explains. A sports dietitian can help set targets for heavy training. **In WeightPlot** Your calorie target comes from your activity level, which already counts weekly workouts. Planned workouts show roughly what they burn, but that number isn't added to your food budget. ## Common Questions ### Should you eat back exercise calories to lose weight? Only if your calorie target leaves exercise out. If your activity level already includes workouts, eating them back counts that exercise twice. ### Should you eat back half of your exercise calories? Half is a common rule of thumb when your target is set to sedentary, but it isn't a research-based figure, so test it against your two-week trend. ### How accurate is an Apple Watch or Fitbit for calories burned? Not very. In a Stanford lab test of seven wrist devices, none estimated calories burned within 20% error. ### Why does eating back exercise calories stall weight loss? Usually the exercise is counted twice, the burn estimate is high, or food is underlogged. Stop adding exercise calories for two weeks and watch the trend. ## The Bottom Line Eating back exercise calories isn't wrong, but it only works when the workout isn't already in your target. Most targets include exercise through the activity level, watch estimates can miss by 20% or more, and the body offsets part of any extra burn. So count exercise once, hold the setup for two weeks and let the trend decide, adjusting in steps of 100 to 200 calories. If you train long or hard, fuel those days properly. The goal isn't a perfect number for every workout, just a setup your weight trend confirms. For more on setting a target and logging accurately, see the [calorie counting guides](/hub/calories-and-counting/). **Key takeaways** - **Count exercise once,** in your activity level or your log. - Wrist devices can miss calories burned by 20% or more. - The body offsets about 28% of extra activity calories. - Test any change against a two-week weight trend. - Long training days need proper fuel. **Sources** - Shcherbina A, et al. Accuracy in wrist-worn, sensor-based measurements of heart rate and energy expenditure. J Pers Med. 2017;7(2):3. - O'Driscoll R, et al. How well do activity monitors estimate energy expenditure? Br J Sports Med. 2020;54:332–340. - Dietary Guidelines Advisory Committee. Appendix E-3.1.A3: Energy levels for USDA Food Patterns. 2015. - Herrmann SD, et al. 2024 Adult Compendium of Physical Activities. J Sport Health Sci. 2024;13:6–12. - Careau V, et al. Energy compensation and adiposity in humans. Curr Biol. 2021;31:4659–4666. - Pontzer H, et al. Constrained total energy expenditure and metabolic adaptation to physical activity. Curr Biol. 2016;26:410–417. - Donnelly JE, et al. Aerobic exercise alone results in clinically significant weight loss: Midwest Exercise Trial 2. Obesity. 2013;21:E219–E228. - Thomas DM, et al. Why do individuals not lose more weight from an exercise intervention? Obes Rev. 2012;13:835–847. - Lichtman SW, et al. Discrepancy between self-reported and actual caloric intake and exercise in obese subjects. N Engl J Med. 1992;327:1893–1898. - Mountjoy M, et al. IOC consensus statement on Relative Energy Deficiency in Sport. Br J Sports Med. 2023;57:1073–1097.