Smartwatches and fitness trackers have become everyday tools for millions of people who want to monitor their exercise, heart rate, steps and calorie expenditure.

But a new study from researchers at Florida International University suggests that users should be cautious when interpreting one of the most popular measurements on these devices: calories burned during exercise.

Researchers tested four widely used fitness watches against laboratory equipment designed to measure energy expenditure and found that the watches frequently overestimated the number of calories participants burned.

The findings raise an important question for people using wearable technology for weight management and fitness: How accurate is the calorie number displayed on your wrist?

Four Popular Fitness Watches Tested

The researchers evaluated four commercially available devices:

• Apple Watch Series 8

• Fitbit Sense 2

• Samsung Galaxy Watch 5

• Garmin Forerunner 955

The devices were tested against a COSMED K5 metabolic analyzer.

The metabolic analyzer is considered a much more precise method for measuring energy expenditure in controlled exercise testing.

The study involved 58 adults who completed controlled cycling sessions.

Participants exercised at different intensities while researchers compared the calorie estimates generated by the watches with measurements from the laboratory equipment.

Watches Often Overestimated Calories

The researchers found that calorie estimates from the watches could differ substantially from the laboratory measurements.

Typical errors were approximately 15% to 25%, although some devices performed worse than others.

Apple's device was the most accurate among the watches tested.

Garmin and Samsung showed some of the largest overestimations.

The Fitbit device presented additional problems during the study, including occasional implausibly low readings and missing measurements.

Because some Fitbit data appeared unreliable, researchers excluded a significant portion of those measurements from their analysis.

The findings do not mean that every calorie estimate from every smartwatch is wrong by the same amount.

Instead, they demonstrate that calorie expenditure is difficult for consumer devices to measure precisely.

Why Are Calories So Difficult to Measure?

Unlike counting steps, estimating calories burned is considerably more complicated.

A smartwatch cannot directly measure how much energy a person's body is using.

Instead, it relies on sensors and algorithms.

Depending on the device, those measurements can include:

• Heart rate

• Movement

• Speed

• Exercise duration

• Body characteristics

• GPS information

• Other physiological signals

The watch then uses an algorithm to estimate energy expenditure.

The problem is that two people performing the same exercise can burn different amounts of energy.

Body composition, fitness level, age, exercise intensity and individual physiology can all influence energy expenditure.

This makes it difficult to create a universal formula that works equally well for everyone.

Body Fat May Affect Accuracy

One particularly interesting finding from the study was the relationship between body fat and measurement error.

Researchers found that the error tended to increase as body fat increased.

The reason is not yet fully understood.

The study's lead author, Jason Kostrna, suggested that the algorithms used by wearable devices may contribute to the problem.

The algorithms are developed using datasets, but consumers generally do not know exactly how those datasets were constructed or how representative they are of the wider population.

If an algorithm performs well for one group of people but less accurately for another, the resulting calorie estimates can become misleading.

Why This Matters for Weight Management

The issue becomes particularly important for people trying to lose weight.

Imagine that a person exercises and believes a smartwatch has recorded 500 calories burned.

If the actual energy expenditure was significantly lower, the person could unintentionally eat more calories than they actually burned.

Over time, repeated inaccuracies could affect weight-management decisions.

This does not mean that fitness trackers are useless.

They can still provide valuable information about activity levels, exercise frequency, heart rate trends and overall behavior.

The key is understanding what the numbers can and cannot tell you.

Fitness Trackers Are Still Useful

A smartwatch can be an excellent motivational tool.

For example, someone who normally walks 3,000 steps per day might use a tracker to gradually increase activity to 6,000 or 8,000 steps.

The exact calorie number is less important in that situation.

The device can help users identify trends.

A person can also use a smartwatch to monitor:

• Daily steps

• Exercise duration

• Heart-rate trends

• Workout frequency

• Running distance

• Cycling distance

• Sleep patterns

• Training consistency

These measurements can help people understand their habits.

The problem arises when users treat an estimated calorie number as if it were a laboratory measurement.

What Should Users Do?

The researchers' findings suggest that people should avoid relying exclusively on smartwatch calorie estimates.

Instead, users can consider several measurements together.

For weight management, long-term changes in body weight and body composition may provide more meaningful information than a single workout's estimated calorie expenditure.

For fitness improvement, performance measurements such as running pace, strength, endurance or exercise consistency can also be more useful.

A smartwatch should therefore be viewed as a tracking and feedback tool, rather than a medical-grade calorie measurement device.

The Technology Could Improve

The findings do not necessarily mean wearable technology has reached a dead end.

On the contrary, studies such as this can help manufacturers improve their algorithms.

Researchers are continuing to test wearable devices under different conditions and with different populations.

The FIU research team is also studying how accurately these devices estimate energy expenditure during strength training.

That could be particularly challenging for wearable algorithms because strength workouts involve short periods of intense effort followed by rest.

Running or cycling at a relatively consistent intensity may be easier to model than resistance training.

A Growing Role for Wearables

Despite their limitations, wearable devices are becoming increasingly important in fitness and preventive health.

Smartwatches can collect enormous amounts of information over long periods.

That creates opportunities for users and researchers to identify patterns that would be difficult to observe manually.

Future devices may combine improved sensors with more sophisticated algorithms and personalized models.

Instead of relying on generalized assumptions, future systems could potentially adapt their estimates to individual users.

However, better technology will still require independent testing.

Consumers should not assume that a sophisticated-looking number is automatically accurate.

The Bigger Lesson

The most important lesson from the study is not that people should stop using fitness trackers.

It is that measurement and estimation are not the same thing.

A smartwatch estimates calorie expenditure using sensor data and algorithms.

Laboratory equipment can measure energy expenditure much more directly under controlled conditions.

For everyday fitness, an estimate can still be useful.

But when a person is making important decisions about diet, exercise or weight management, it is better to look at broader trends rather than treating one number on a smartwatch screen as absolute truth.

Wearable technology will probably continue to improve.

As algorithms become more personalized and sensors become more advanced, calorie estimates may become more reliable.

Until then, users should treat calorie-burn figures as estimates—and use their fitness trackers primarily as tools for monitoring habits, activity and progress.

Source: Florida International University, September 2026; study published in PLOS One.