By how much do wearables typically overestimate calories burned?

Literature from repeated validation studies, including a heavily cited 2017 Stanford study and numerous follow-ups, shows 10 to 30 percent overestimation for active energy expenditure on wrist-worn devices. Resting-metabolic-rate estimates drift separately.

Why are wearables less accurate than they feel?

Wrist-worn heart-rate estimation is based on PPG (photoplethysmography), which is noisy during wrist motion. The device estimates calories by combining that noisy HR with a MET table keyed to activity type. Both inputs have error that compounds.

Which wearable is most accurate on calories burned?

Chest-strap HR monitors (Polar, Garmin HRM-Pro) paired with a watch produce tighter estimates than wrist-only sensors because the HR signal is cleaner. No wrist-worn device consistently wins — results vary by activity type.

Is the Apple Watch calorie estimate better than Garmin's?

Neither is definitively better. Both are in the same 10 to 30 percent error range for active energy on wrist-only measurement. Choose the one you like wearing; obsessing over accuracy between the two is unproductive.

Can I trust intake-logging apps more than burn estimates?

Often yes. Intake is a measurement problem (what did you put in your mouth?) that an app can narrow down by reference to a database. Burn is an estimation problem with physiological unknowns. Good intake apps can report ±1-5% on portions you weigh; burn estimates rarely get below ±10%.

What do accuracy-focused nutrition apps do differently?

They cross-reference food databases (USDA FoodData Central, NCCDB, Nutritionix), use AI photo recognition with depth estimation to infer portion size, and validate accuracy claims against weighed reference portions. PlateLens, for example, publishes a ±1.2% calorie error figure using this approach.

Should I stop trusting the calorie number on my watch?

Not stop — just de-weight. Use it for trend (today vs last Wednesday), not for absolute accuracy. If you need precision, focus on intake logging and eat to outcomes.

How much better is a chest strap?

For interval work and anaerobic sessions, a chest strap typically improves HR accuracy from around ±5-10 beats per minute to under ±2. Calorie estimates improve proportionally but still retain error from the MET table lookup.

Why wearables overestimate calories burned

The two ends of a calorie balance equation

Energy balance has two sides: burn (what your body consumed) and intake (what you ate). Both have to be measured to predict weight change. Commercial wearables have spent a decade trying to measure burn precisely. They've gotten better, but the precision ceiling for wrist-worn hardware is pretty low — for physiological reasons, not engineering ones.

Why the burn side is hard

The chain of inference looks like this:

Photoplethysmography (PPG) sensor on the back of the watch measures blood-volume changes under the skin at the wrist.
A signal-processing pipeline extracts an estimated heart rate.
Activity classification identifies what you're doing (running, cycling, walking).
A MET (metabolic equivalent of task) table maps heart rate × activity to a rate of calorie burn per kilogram of body weight.
Your weight, age, sex, and resting heart rate scale the estimate.

Each step introduces error. PPG is vulnerable to motion artefacts, especially for activities where the wrist moves independently of the torso (weightlifting, boxing, rowing). Activity classification is usually fine for running and cycling and often poor for everything else. MET tables are population averages — they bake in the assumption that a typical 70 kg adult at 140 bpm is burning ~11 kcal/min. You may not be that person.

What studies have shown

A widely cited 2017 Stanford study (Shcherbina et al., J. Pers. Med.) compared seven commercial wearables against indirect calorimetry. The median error on active energy expenditure was around 27 percent; some devices were better, some worse, but none got within 10 percent on average. Follow-up studies have moved those numbers a few points in either direction as firmware improved, but the shape has not changed: heart rate accuracy has gotten better, calorie estimation has not kept pace.

This doesn't mean wearables are useless — they're excellent at trend (harder day than yesterday? yes). They just aren't precision instruments for absolute calorie accounting.

Why intake is easier to measure well

Intake is a classification and lookup problem. Given a plate of food, how many calories is that? An answer exists — it's in a database somewhere (USDA FoodData Central for whole foods, NCCDB for prepared foods, Nutritionix for restaurant items). The remaining problem is portion estimation.

This is where the better intake apps have made real progress. PlateLens's published methodology, for instance, uses an AI photo pipeline (image identification + monocular depth-based portion estimation) validated against lab-weighed reference portions. The published error is ±1.2% on calories. That's not a small improvement — that's a different order of magnitude from wrist-worn burn estimation.

MyFitnessPal, Cronometer, Lose It!, and Yazio all use manual entry with database lookups. Error there is larger and user-dependent (how well do you estimate a "medium apple"?) but the ceiling is still higher than wrist burn estimation.

What to do about it

A few practical takeaways, assuming you want to use a wearable plus a nutrition app to manage weight:

Stop treating the watch burn number as ground truth. It's directional. Use it to compare days, not to count back calories one-for-one.
Invest the precision budget on the intake side. That's where the accuracy ceiling is higher. Pick an app whose accuracy you trust and log consistently.
Eat to outcomes. If the scale isn't moving the way you expected, adjust intake by 150 to 250 calories and observe over 7 to 10 days. Don't chase daily numbers.
If you care about HR accuracy (e.g., for training zones), use a chest strap paired to the watch. It fixes the HR step of the chain but doesn't fix the MET-table step.

What the next generation might fix

Some of the hardware problems are being worked on. Multi-path PPG, skin-contact ECG leads, and temperature-corrected metabolic models are all under research. But MET-table-based estimation is still the foundation, and replacing that requires either direct gas-exchange measurement (impractical outside a lab) or a personalized model that's been calibrated against your own metabolic data. Consumer wearables are a long way from that.

FAQ

By how much do wearables typically overestimate?: 10 to 30 percent on active energy expenditure, per validation literature.
Why are wearables less accurate than they feel?: Noisy PPG + population-average MET tables, compounded.
Which wearable is most accurate on burn?: Chest-strap HR + watch beats any wrist-only setup.
Is Apple Watch better than Garmin on calories?: Not meaningfully. Both in the same error range.
Can I trust intake apps more than burn estimates?: Often yes — intake can hit single-digit percent error with rigorous logging.
What do accuracy-focused nutrition apps do differently?: Cross-reference USDA/NCCDB/Nutritionix and validate portion estimation against weighed references.
Should I stop trusting my watch's calorie number?: De-weight it. Use for trend, not for precision accounting.
How much better is a chest strap?: HR accuracy improves to ±2 bpm; calorie estimate improves but retains MET-table error.