Which Numbers on Your Fitness Tracker Are Real

Your watch says you burned 612 calories on that run. You log it, eat a little more at dinner because you earned it, and close the rings. The loop feels scientific.

It is not, quite. Across the three headline metrics, consumer fitness trackers average about 67% accuracy. That is not useless. But it is a long way from the precision the interface implies, and the metric people act on most is the one the research trusts least.

Knowing which numbers on your fitness tracker hold up, and which ones are educated guesses, changes how you should use the thing on your wrist.

Key Takeaways

• Cumulative accuracy across heart rate, steps and calories averages 67.4%, ranging from 62% to 74% by device
• Heart rate is the most reliable metric, with the Apple Watch measured at 86.3%
• Calorie burn is the weakest, around 71% even on the best performer, and it is the number most people act on
• Among major brands, only the Apple Watch holds FDA clearance for ECG-based atrial fibrillation detection

What fitness tracker accuracy actually measures

A systematic review and meta-analysis published through the National Library of Medicine pooled real-world accuracy studies across the main consumer devices. The cumulative figure for heart rate, step count and energy expenditure landed between 62.09% and 73.53%, averaging 67.40%.

Accuracy here means agreement with a laboratory reference, not whether the device is broken. A tracker reading 8% high every single day is inaccurate but perfectly consistent, and consistency is what makes trend data useful.

So the honest framing is this. Your tracker is a decent instrument for comparing you to yourself over weeks. It is a poor instrument for comparing today's absolute number to anything.

Heart rate is the number you can trust

Of the three metrics, optical heart rate holds up best. The Apple Watch measures at 86.31% accuracy, the strongest result in the category.

The mechanism is simple enough. A green LED shines into your wrist and a sensor reads how much light comes back as blood pulses through. Blood absorbs green light, so the reflection changes rhythmically, and the device counts the rhythm.

That works well at rest and degrades under movement. Some devices lose up to 30% of their accuracy during activity, because a shifting wrist band, sweat and tendon movement all corrupt the signal. Interval work is the worst case: rapid changes are exactly what optical sensors lag on.

Practical version: trust your resting heart rate, treat mid-interval readings as approximate. If you train by heart rate zones seriously, a chest strap remains meaningfully better, because it measures electrical activity rather than inferring from light.

Step count is good enough, with one blind spot

Steps are the oldest metric and the best understood. Garmin leads at 82.58% accuracy, with the rest of the field clustered not far behind.

The blind spot is what counts as a step. Wrist-worn accelerometers detect arm swing, so pushing a pram, carrying shopping or holding a handrail on a treadmill can lose you hundreds of genuine steps. The reverse also happens. Washing up, folding laundry and gesturing during a phone call all generate phantom steps.

For most people these roughly cancel out over a week, which is why step counts work fine as a habit signal and badly as a precise measurement.

If you want to test your own device, count 100 steps deliberately and compare. Most trackers land within five either way. A device that misses by fifteen is telling you something about its placement, not about your walking.

Calories are the number to stop believing

Energy expenditure is where the gap between confidence and accuracy is widest. Even the Apple Watch, the strongest performer, sits at 71.02%.

The reason is structural. No wrist device measures energy burn. It estimates it, from heart rate, motion, and the age, weight, height and sex you typed in at setup. Every one of those inputs carries error, and they multiply.

Which would not matter much, except this is the number that drives behaviour.

Run the arithmetic. Suppose your tracker credits you with 600 calories from a workout and the true figure is 450, a 25% overstatement well inside the observed error band. If you eat those 150 phantom calories back, that is 54,750 calories a year. At roughly 3,500 calories per pound of body fat, you have accounted for about 15 pounds of drift you cannot explain.

That is the single most common way a tracker makes someone's goal harder rather than easier. The device did nothing wrong. It offered an estimate and the interface presented it as a measurement.

The fix is to stop treating calorie burn as a budget you can spend. Use it as a relative signal: today was harder than Tuesday. That comparison is valid even when the absolute number is not.

This is the same problem covered in what biological age tests actually measure, where a precise-looking output hides a wide confidence interval. Our piece on why sleep outranks diet and exercise covers the metric these devices are arguably best at capturing.

What FDA clearance does and does not mean

There is one line worth knowing, because marketing blurs it constantly.

The Apple Watch holds FDA clearance for ECG-based atrial fibrillation detection in adults. As of June, no Fitbit, Garmin, Polar or Coros device held FDA clearance for any cardiac diagnostic function.

Those brands still show heart data, and that data is still useful. It has simply not been through the regulatory process that tests a specific diagnostic claim, and the meta-analysis is blunt that wearable accuracy as a disease detection tool remains uncertain.

Read the wording on any health alert accordingly. "May indicate" is doing real work in those sentences. A wearable alert is a prompt to see a clinician, never a result.

How to get real value from the device you own

Four habits turn a 67% instrument into something genuinely useful.

Compare yourself to yourself. Your resting heart rate over eight weeks is a strong signal. Your resting heart rate against your friend's is noise, because the two devices have different error profiles.

Wear it the same way every day. One finger width above the wrist bone, snug enough not to slide. Placement changes the optical reading more than most people realise, and inconsistent placement destroys the trend data that is the whole point.

Never eat back calorie burn. If you are tracking intake, set it from your goal and treat exercise as unbudgeted.

Watch the direction, not the digit. Resting heart rate creeping up across a fortnight is worth attention whether the absolute number is right or not.

Frequently Asked Questions

How accurate are fitness trackers?

About 67% on average across heart rate, step count and calorie burn combined, with devices ranging from 62% to 74%. Accuracy varies sharply by metric, so a single overall figure hides the fact that heart rate is far more reliable than calorie estimates.

Which fitness tracker is the most accurate?

It depends on the metric. The Apple Watch leads on heart rate at 86.3%, Garmin leads on step count at 82.6%, and no device performs well on energy expenditure, where even the best sits near 71%.

Are fitness tracker calorie counts accurate?

No. Calorie burn is the least reliable metric, around 71% at best, because no wrist device measures energy expenditure directly. It estimates from heart rate, motion and the profile data you entered, and those errors compound.

Can a smartwatch detect heart problems?

Only within limits. The Apple Watch has FDA clearance for ECG-based atrial fibrillation detection, while major rivals hold no FDA clearance for cardiac diagnostics. Any alert is a reason to see a doctor, not a diagnosis.

Should I use a chest strap instead of a wrist tracker?

For structured heart rate training, yes. Chest straps read electrical activity rather than inferring from light, so they hold accuracy during intervals where wrist sensors can lose up to 30%. For daily activity tracking a wrist device is fine.

The short version

Your tracker is a good habit instrument wearing the costume of a laboratory device. Heart rate is solid, steps are close enough, and calorie burn is an estimate that the interface presents with unearned confidence.

Use the trend and ignore the decimal place. Never spend the calories it says you earned. And treat any health alert as a reason to book an appointment rather than a result to act on alone.

The most valuable thing on your wrist is not the number. It is the eight-week line underneath it.