How Accurate Is Apple Watch Calorie Tracking? (What the Research Says)

Short answer: your Apple Watch measures heart rate very well and calories quite badly. Independent validation work puts wrist-wearable energy expenditure error somewhere in the 20-40% range, and Apple Watch is generally at the better end of that spread. That means a workout the watch calls 500 calories could plausibly be 350 or 650. The good news: it's usually wrong in a consistent direction, which makes it genuinely useful as a relative signal even though it's a poor measurement.

This is one of the most-searched questions about the device, and most answers are either "it's fine" or "it's useless." Neither is right, and the useful answer is about which parts are accurate.

The Two Things It Measures, and Why Only One Works

Your watch does not measure calories. It infers them, and the inference chain has a weak link.

What it measures directly: heart rate, via green LEDs and photodiodes reading blood volume changes at your wrist. Motion, via accelerometer and gyroscope. Location and pace, via GPS. This part is solid. Stanford's well-known evaluation of six wrist wearables found heart rate error under about 5% for most devices in most conditions.

What it estimates: energy expenditure. To get there, the watch takes your heart rate and motion, adds your age, sex, height, weight, and runs it through a model of how much oxygen a person like you consumes at that heart rate. Every step of that chain adds error.

The same Stanford work found the gap starkly: none of the six devices tested got energy expenditure error below 20%, even though their heart rate readings were good. Apple Watch had the lowest error of the group. Later work has broadly confirmed the pattern - good pulse, poor calories.

The reason is a physiological one. Heart rate is a proxy for oxygen consumption, not a measurement of it, and the relationship between the two varies enormously with fitness, hydration, temperature, caffeine, stress, and the type of contraction your muscles are doing.

Why the heart-rate-to-calories link breaks

Four specific mechanisms pull the two apart, and knowing them tells you exactly when to distrust the number.

Cardiac drift. Hold a steady effort for 45 minutes and your heart rate climbs anyway, by 5-10% or more, as blood volume falls and your body diverts blood to the skin for cooling. Your oxygen consumption hasn't changed. The watch reads the drift as increasing effort and keeps adding calories that you never burned.

Thermoregulation. Exercise in heat and your heart rate runs substantially higher at any given pace. The extra work is pumping blood to your skin, not to your muscles. A summer run and a winter run at identical pace produce very different calorie figures for identical energy cost.

Contraction type. Isometric and heavy concentric work - a plank hold, a heavy set of deadlifts - drives blood pressure and heart rate up sharply while consuming comparatively little oxygen. The model has no way to distinguish this from running, so it treats a heart rate of 160 identically either way.

Individual variation in fitness. Two people at 150 bpm can be consuming very different amounts of oxygen, because a trained heart moves far more blood per beat. The watch tries to correct for this using your recent workout data, and it does so imperfectly, which is why the error is bigger for people at the extremes of fitness.

None of these are bugs. They are the unavoidable consequence of estimating one thing from a different thing, and no wrist device escapes them.

Where It's Most Accurate, and Where It Isn't

Activity	Accuracy	Why
Outdoor walk with GPS	Best	Steady state, GPS-verified pace, distance-based model
Outdoor run with GPS	Best	Same reasons, plus the calibration data is built around it
Steady cycling	Good	Consistent effort, though wrist HR suffers from grip tension
Rowing, elliptical	Moderate	Arm motion confuses the accelerometer
Hiking	Moderate	Elevation helps, but pack weight is invisible to the watch
HIIT	Poor	Heart rate lags the effort in both directions
Strength training	Worst	High HR, low oxygen cost. The model badly overestimates.
Yoga, pilates	Poor	Isometric holds spike heart rate with little energy cost

Strength training is the single largest error source. A heavy set of squats drives your heart rate to 160, and the algorithm reads that as running-level oxygen consumption. It isn't. The actual energy cost of resistance training is far lower than the heart rate implies, and Apple Watch routinely overestimates lifting sessions substantially.

Caffeine, heat, dehydration, and stress all push heart rate up without pushing energy expenditure up, and all of them therefore inflate your calorie number.

The strength training problem, in numbers

Worth spelling out because it's the largest single error and the one people most often build a diet around.

A 45-minute resistance session for an average adult genuinely costs somewhere in the region of 150-250 calories. Apple Watch commonly reports 400-600 for the same session. The gap comes from time under tension: in a 45-minute workout you might do 20 working sets of 40 seconds each, which is 13 minutes of actual work spread across 45 minutes of elevated heart rate. The watch bills you for all 45.

The usual counter-argument is EPOC - the "afterburn effect," where your metabolism stays elevated post-workout. It is real, and it is small. Research generally puts it at 6-15% of the session's cost, so a genuine 200-calorie workout buys you perhaps another 20-30 calories over the following hours. It does not close a 300-calorie gap, and it is already partly reflected in your watch's elevated post-workout readings anyway.

None of this is an argument against lifting. Resistance training is the single best thing you can do to preserve muscle in a deficit, which is the whole mechanism behind body recomposition. It's an argument against eating as though lifting burned what your watch says it did.

Active vs Total Calories: The Confusion Worth Clearing Up

This trips up almost everyone, and getting it wrong costs you hundreds of calories a day.

• Active calories = energy burned through movement, on top of simply being alive.
• Resting calories = energy your body uses to keep you running: heart, brain, breathing, temperature.
• Total calories = active + resting.

The Move ring tracks active calories only. The Total figure inside a workout summary includes the resting burn you'd have had anyway while sitting on the sofa.

Which one goes into your food tracker? Almost always active only. Apps like MyFitnessPal already build your resting burn into your daily target. Logging total calories double-counts your resting metabolism and can silently add 1,500-2,000 phantom calories a day to your allowance. If you want a sense of what your resting burn actually is, the REE calculator will estimate it.

A quick sanity check on the scale of that mistake:

	Active only	Total (active + resting)
Typical workout summary	320 kcal	380 kcal
Typical full day	550 kcal	2,150 kcal
Added to a 1,800 kcal target	2,350 kcal	3,950 kcal

The second column of that last row is how a determined person eats at a surplus while believing they are dieting.

The Move ring is not a calorie budget

A related trap. The Move ring shows active calories against a goal you set, and the goal is arbitrary - you picked it, or Apple suggested it. Closing it means you hit an activity target. It says nothing about whether you're in a deficit, and it isn't calibrated to your body or your diet.

Two consequences worth internalising. A closed ring on a day you ate 3,000 calories is still a surplus. And chasing the ring late in the evening is the most common way people convert a useful habit into a compulsive one. Use it as an activity nudge, which is what it's good at.

How to Make It as Accurate as It Can Be

You can't fix the model, but you can stop feeding it bad inputs. In rough order of impact:

1. Fix your Health profile. Age, sex, height, and especially weight. Weight is a direct multiplier in the calculation, and a figure that's 10 kg stale skews every session. Health → Profile → Health Details. Update it whenever your weight changes meaningfully.

2. Calibrate the watch. Take your iPhone, go outside somewhere flat and open, and do 20 minutes of Outdoor Walk and 20 minutes of Outdoor Run with clear GPS. The watch compares its motion data against GPS-verified distance and builds a personal stride model. Recalibrate after significant weight change or a big shift in fitness. Reset the existing data in Watch app → Privacy → Reset Fitness Calibration Data.

3. Wear it properly. Snug, one finger-width above the wrist bone, not sliding around. A loose band is the number one cause of bad optical heart rate, and bad heart rate means bad calories. Tighten it during workouts specifically.

4. Always start a Workout. Background tracking is much cruder. Starting a workout switches to continuous heart rate sampling and an activity-specific model.

5. Pick the right workout type. "Functional Strength Training" and "Outdoor Run" use different models. Choosing "Other" gives you a generic estimate that's worse than either.

6. Clean, dry skin. Sweat, sunscreen, lotion, and tattoos over the sensor all degrade the optical reading. Dark ink is a genuine and well-documented problem.

7. Set the wheelchair and other profile options correctly. Apple has a separate model for wheelchair users, with pushes rather than steps. If it applies to you and isn't set, every number is wrong.

Doing all seven typically pulls the error in, but it won't get you to accurate. It gets you to consistently wrong, which is what you actually want.

What resetting calibration actually does

Worth understanding before you do it. Your watch builds a personal model of your stride - how far you travel per step at each cadence - by comparing accelerometer data against GPS-verified distance during outdoor workouts. That model drives distance estimates, which drive calorie estimates when GPS isn't available.

Resetting it (Watch app → Privacy → Reset Fitness Calibration Data) throws that model away and starts over. Useful after a significant weight change, a major fitness change, or if you suspect the data got polluted - by treadmill workouts logged as outdoor runs, for example. Not useful as a routine fix, since you lose a working model and spend weeks rebuilding it.

After a reset, do 20 minutes of outdoor walking and 20 minutes of outdoor running with GPS and your iPhone, on flat open ground, to rebuild it quickly.

Why "Consistently Wrong" Is Good Enough

Here's the reframe that makes the whole question tractable.

If your watch overestimates by 25% every single day, then a day it calls 600 active calories is genuinely bigger than a day it calls 400. The ratio holds even though neither number is true. You can use it to compare Tuesday to Wednesday, to see whether this month was more active than last, and to notice when your training volume drops off.

What breaks is using the number as an input to a calorie budget. That's where a 25% error becomes 400 real calories a day, and a deficit you believe in becomes maintenance.

Here's how that compounds over a single week, using a person targeting a 500-calorie daily deficit who eats back everything the watch reports:

Day	Watch says burned	Actually burned	Phantom calories
Mon (lifting)	480	220	260
Tue (walk)	310	280	30
Wed (HIIT)	520	350	170
Thu (rest)	180	160	20
Fri (lifting)	460	210	250
Sat (long walk)	620	560	60
Sun (rest)	200	180	20
Week	2,770	1,960	810

That's 810 calories of a planned 3,500-calorie weekly deficit gone - roughly a quarter - purely from trusting the estimate. Notice where it concentrates: the two lifting days account for almost two-thirds of the error. Someone doing five gym sessions a week could erase their deficit entirely.

The eat-back problem

If you eat back everything your watch says you burned, you will very likely stall. The usual advice, and it's sound:

• Eat back around 50% of active calories, as a starting point.
• Watch your weight trend for 3-4 weeks.
• Adjust from what the scale actually did, not from what the watch claimed.

Your weight trend is the only honest arbiter here. If you're eating back half and losing 0.5 kg a week, the arithmetic is working regardless of whether the watch's numbers were true. If you're eating back half and not losing, eat back less. This is also the most common explanation behind being in a calorie deficit but gaining weight - the deficit was never as large as the app suggested.

A four-week protocol to calibrate against reality

You can work out your own personal correction factor without knowing any true number. This is the single most useful thing in this article.

Weeks 1-2: establish the baseline. Eat at a fixed calorie target every day. Don't adjust for workouts. Weigh yourself every morning under the same conditions and take the weekly average, not individual days. Log your watch's active calories but don't act on them.

Weeks 3-4: read the result. Compare your week-2 average weight against your week-4 average. Roughly 7,700 calories corresponds to 1 kg of body mass, so:

• Lost 0.5 kg/week → you were running about a 550 kcal/day deficit. Your intake target was right.
• Weight flat → you were at maintenance. Your true maintenance is your intake, whatever the watch claimed.
• Gained → you were in a surplus. Reduce intake, not the watch's number.

Then, if you want a correction factor: compare the deficit the arithmetic implies against what your watch and your food log jointly predicted. If the watch predicted a 900 kcal/day deficit and reality delivered 550, your personal factor is about 0.6 - eat back roughly 60% of what it reports.

Two honest caveats. Your food logging is also wrong, typically underestimating by 20-30%, so this factor absorbs both errors at once. That's fine, because you want a factor that works in practice, not one that's physiologically pure. And four weeks is the minimum, because water weight, glycogen, sodium, and menstrual cycle noise all swamp the signal over shorter windows. Our guide on why weight fluctuates covers why the daily scale reading is so unreliable.

How It Compares to Other Devices

Device	Heart rate	Calorie estimate	Notes
Apple Watch	Very good	Best of the wrist wearables, still 20%+ error	Strongest on GPS walk and run
Garmin	Very good	Similar range, Firstbeat model	Better for structured endurance training
Fitbit	Good	Similar range	Historically ran high
WHOOP	Very good	Similar range	Strain model rather than a raw calorie figure
Chest strap	Excellent	Still an estimate	Better HR input, same modelling problem
Metabolic cart	n/a	The actual measurement	Lab only, measures gas exchange directly

The point of the table: no wrist device solves this, because the limitation isn't the sensor. A chest strap gives you a better heart rate and roughly the same calorie error, because heart rate was never a clean proxy for oxygen consumption in the first place. Only measuring the gas you exhale gets you a real number.

What to Actually Do With the Number

Use it for: comparing your own days and weeks, spotting when activity drops off, judging relative session intensity, and driving activity goals.

Don't use it for: setting a food budget, comparing yourself to someone else, or justifying a meal.

Anchor everything to your weight trend. Set a calorie target from a calculation, hold it steady for three or four weeks, and let the trend tell you whether the number was right. That loop works whatever your watch says.

The two errors cancel more often than you'd think

One reassuring point to end on. Your watch overestimates your burn, typically by 20-40%. Your food log underestimates your intake, typically by 20-30% - unlogged oils, eyeballed portions, the handful of nuts, the milk in three coffees.

Those errors run in opposite directions, and for many people they partially cancel. That's why calorie tracking works at all despite both halves being unreliable. It also means that fixing only one side - suddenly logging food meticulously while still trusting your watch - can make your results worse before they get better.

The way out isn't more precision on either side. It's anchoring both to the scale, holding a target steady long enough to read the trend, and adjusting the target rather than the estimates.

Put the Number Next to Its Consequences

A calorie estimate on its own is a number with no feedback loop. It only becomes useful when you can see it against what you ate and what your weight actually did, over enough weeks for the noise to cancel out.

That's what Bodly is built around. Your Apple Watch's active calories sit next to your logged food, your macros, your weight trend, and your body measurements - so instead of trusting the watch, you can check it. If your watch says you burned 600 and you ate 2,200 and your weight hasn't moved in a month, the answer is right there, and it's a far more reliable guide than any single estimate.

Try Bodly if you'd rather calibrate against reality. For the deficit-that-isn't problem, see calorie deficit but gaining weight, and for tracking apps generally, best free MyFitnessPal alternatives.

Bottom line: the heart rate is trustworthy, the calorie number is not. Expect roughly 20-40% error, log active calories rather than total, eat back about half, and let your weight trend over a month be the thing you actually believe.

Keep reading

If the calorie figure is the part you cannot trust, the answer is usually to anchor it to something measurable. The best free weight and weight loss tracker apps compares apps that reconcile intake against an actual weight trend instead of taking your watch at its word.

Sources

• Journal of Personalized Medicine: Accuracy in Wrist-Worn, Sensor-Based Measurements of Heart Rate and Energy Expenditure in a Diverse Cohort (Stanford)
• Apple Support: Get the most accurate measurements using your Apple Watch
• Apple Support: Calibrate your Apple Watch for improved Workout and Activity accuracy
• JMIR mHealth and uHealth: Accuracy of Wearable Devices for Estimating Total Energy Expenditure