Should You Eat Back Your Exercise Calories? A Practical Answer With Numbers

TL;DR

• It depends on how your target was set. If you used an activity multiplier that already counts your training, eating back exercise calories double-counts it.
• Wearables overestimate calorie burn for most activities. Independent studies put the error at roughly 25-90%, with walking the most accurate and strength training among the worst.
• If your target is set at sedentary and you train on top, eat back about half of what your watch reports. That rule covers most of the error.
• On very long or very hard sessions (90 minutes plus), eat more back, mostly as carbs. Under-fuelling hurts performance and recovery more than a slightly smaller deficit helps.
• Two to three weeks of scale trend data beats any formula. If weight is dropping at your target rate, whatever you are doing is correct.

Start with how your target was calculated

This is the question most people skip, and it decides the whole answer.

There are two common ways to set a calorie target:

1. Activity multiplier. You take your BMR and multiply it by something like 1.375 (lightly active) or 1.55 (moderately active). Most online calculators, including the method in our TDEE guide, work this way. The multiplier already includes your typical training.
2. Sedentary base plus exercise. You set your target as if you did nothing (multiplier around 1.2), then add exercise on top each day you train. This is how many apps work when they link to a watch.

If you used method 1, your training is already in the number. Eating back the calories your watch reports means counting the same session twice, and a 400 kcal session counted twice, four times a week, wipes out most of a 500 kcal daily deficit.

If you used method 2, you do need to eat something back, otherwise a hard training day becomes a 1,000 kcal deficit and you will feel it.

Quick check

Look at your target. If it changes on days you train, you are on method 2. If it is the same every day and you told the calculator you exercise, you are on method 1. Do not eat back.

Why your watch is not a calorie meter

Wearables measure heart rate and movement, then estimate energy burn with a model. Heart rate is a reasonable signal for steady cardio. It is a poor one for lifting, where heart rate spikes from effort and breath-holding without a matching calorie burn.

The research is fairly consistent:

• Independent validation studies of wrist-worn devices consistently find heart rate tracking is fairly accurate, usually within a few percent, while energy expenditure estimates are far less reliable, with errors commonly ranging from roughly 25% for the best-performing devices to well over 80% for the worst.
• Errors run high far more often than low. Few people find their watch underestimates.
• Accuracy varies by person. Two people doing the same session on the same model can see very different errors.

On top of that, the number on your watch is usually gross burn: everything you burned during the session, including the 80-100 kcal an hour you would have burned sitting on the sofa. The extra you burned because you exercised is smaller than the figure shown.

What that means in practice

A 60-minute weights session showing 450 kcal might be 200-300 kcal of real extra burn. A 45-minute run showing 500 kcal might be 350-450. The direction of the error is predictable even if the size is not.

The half rule, and when to break it

If your target is set at sedentary and you add exercise on top, a simple rule works for most people: eat back about 50% of what your device reports.

It roughly cancels the typical overestimate and the gross-versus-net gap without you needing to know your own error. It is not precise, and it does not need to be. The scale will tell you if it is wrong.

Eat back more when

• The session is long. Runs, rides or Hyrox sessions over 90 minutes burn enough glycogen that a large deficit on top shows up in the next session. Eat back 60-80%, mostly as carbohydrate. Our runners and endurance guide covers fuelling around long sessions.
• You are already lean or close to maintenance. A smaller body has less margin. Aim for a deficit of no more than about 0.5-0.7% of bodyweight per week.
• Performance is dropping. If lifts stall for two or three weeks, sleep gets worse, or you feel flat in every session, you are probably under-fuelled.

Eat back less (or nothing) when

• The session is short or easy. A 30-minute walk adds maybe 80-120 kcal net. Not worth logging separately if your target already allows for daily movement.
• It is strength training. Watches are least accurate here. Treat lifting as a protein and recovery cue, not a calorie credit.
• Fat loss has stalled. If the trend has been flat for three weeks and you have been eating back, stop eating back before you cut food.

Let the scale settle the argument

Formulas get you to a sensible starting point. Your own data decides from there.

Run whichever approach you choose for two to three weeks without changing it. Weigh daily or several times a week and look at the weekly average, not single days, because water swings of 0.5-1.5 kg are normal. Our guide on why the scale goes up in a deficit explains the noise.

Then compare the trend to your goal:

• Losing 0.5-1% of bodyweight a week: you have it right. Keep going.
• Losing faster than 1% a week and feeling drained: eat back more, or raise the base target by 150-200 kcal.
• Losing nothing for three weeks: eat back less, or lower the base target by 150-200 kcal.

This works because the scale measures what actually happened, including every error in your watch, your logging and your TDEE estimate. It is the only number that is not an estimate.

A simpler alternative: one target for every day

Plenty of people find the daily up-and-down of eating back calories more hassle than it is worth. The alternative is to fold training into one fixed target:

1. Estimate your average weekly training burn using the half rule.
2. Divide by seven and add it to your sedentary target.
3. Eat the same calories every day, and shift some carbs towards training days if you like.

You end up with one number to hit, no reliance on your watch’s daily guess, and an easy test of whether it is working. If you want more flexibility across the week, our piece on carb cycling covers moving carbs around without changing the weekly total.

The short version

Find out whether your target already includes exercise. If it does, do not eat back. If it does not, eat back about half of what your watch says, more on long sessions and less on lifting. Then give it two to three weeks and let the scale trend tell you whether to adjust. Log the food you eat accurately and the exercise number matters far less than it seems.