Fitness apps and wearables use step counts as the primary measure of daily movement. Step detection is generally accurate -- accelerometers are good at counting discrete footstep impacts. The problem is not the count; it is what the count implies.
When a device converts your step count to distance, calories, and activity minutes, it applies an assumed stride length. That assumption is usually based on population averages. If your stride length deviates significantly from the assumed average -- which is true for anyone who is notably shorter or taller than average -- every derived metric inherits a systematic error.
What stride length actually is and how it varies
Stride length is the distance traveled between two consecutive footfalls of the same foot. For walking at a moderate pace, stride length scales reliably with height. The widely cited formulas from biomechanics research:
- Women: stride length (m) = height (m) x 0.413
- Men: stride length (m) = height (m) x 0.415
For running, stride length is typically 30 to 45 percent longer than walking stride at easy-to-moderate pace, and can be 50 to 70 percent longer at faster race paces.
These formulas explain why two people of different heights, walking the same route, take different numbers of steps. They also explain why a device using a single assumed stride for everyone produces different accuracy levels for different users.
The magnitude of the error by height group
Applying the formula to the height range produces a concrete picture of how much the error varies:
At 5 feet 0 (152 cm), the formula gives a walking stride of about 62.7 cm, or roughly 2.06 feet per step. This person takes approximately 2,561 steps per mile when walking.
At 5 feet 7 (170 cm), the average height the device assumes, the formula gives a stride of about 70.2 cm, or 2.30 feet per step, corresponding to approximately 2,294 steps per mile.
At 6 feet 2 (188 cm), the formula gives a stride of about 77.6 cm, or 2.55 feet per step, corresponding to approximately 2,074 steps per mile.
The difference between the shortest and tallest person in this example is 487 steps per mile -- a 23 percent relative difference. Over a 5-mile walk:
- The 5-foot person takes approximately 12,805 steps
- The 5-foot-7 person takes approximately 11,470 steps
- The 6-foot-2 person takes approximately 10,370 steps
A device using the average-height assumption (2,294 steps per mile) will estimate the 5-foot person's 5-mile walk as 4.4 miles and the 6-foot-2 person's 5-mile walk as 5.6 miles. The shorter person's distance is understated; the taller person's is overstated.
How this error propagates to calorie estimates
Calorie burn during walking is a function of distance and body weight, not step count. The standard estimate uses a metabolic equivalent (MET) approach: calories per minute = MET value x body weight in kg / 60.
For brisk walking (MET approximately 4.3), a 70 kg person burns approximately 5.0 calories per minute. At 3.5 mph, 5 miles takes about 86 minutes, burning approximately 430 calories.
When the device gets the distance wrong -- say it estimates 4.4 miles instead of 5 miles for the 5-foot person -- the calorie estimate scales with the distance error: the device reports approximately 378 calories instead of 430, a 12 percent underestimate.
For a calorie-tracking application (weight management, macronutrient planning), this error is significant. The American Heart Association recommends 150 minutes of moderate aerobic activity per week for cardiovascular health; tracking this in calories or distance rather than steps makes the recommended volume more meaningful.
Activity type: the compounding error
Running strides are longer than walking strides by a factor that depends on pace. For an easy jogging pace (9-10 minute miles), stride length is roughly 35 percent longer than walking stride. For a faster tempo pace (7-8 minute miles), stride length may be 50 to 60 percent longer.
A device that does not accurately detect activity type -- or that applies a conservative stride estimate to all activity -- will undercount distance during running. For a 5-foot-8 runner with a walking stride of 69 cm, a running stride at easy pace is approximately 93 cm. Applying the walking stride to a 3-mile run understates the distance by approximately 26 percent (3 miles becomes 2.4 miles in the device's estimate).
Modern accelerometers use machine learning to classify activity type and adjust stride estimates accordingly. However, the classification is imperfect, particularly at activity transitions and during non-standard movements. GPS-based tracking eliminates this problem for outdoor activities by measuring distance directly rather than inferring it from steps.
Using a personalized calculator as a reference
The EvvyTools Steps to Distance Calculator allows a user to input their specific height, gender, and activity type to get a personalized step-to-distance conversion factor. This is not a replacement for GPS tracking on outdoor activity -- GPS is more accurate -- but it is substantially more accurate than a device using a population average stride for all users.
The key output to take from the calculator is the steps-per-mile ratio for walking and running separately. These ratios can be applied to raw step counts from any device to produce a more accurate distance estimate:
- If your walking steps-per-mile is 2,400 and your device reports 9,600 steps on a walk, your estimated distance is 4.0 miles.
- If your running steps-per-mile is 1,700 and you ran 5,100 steps, your estimated distance is 3.0 miles.
The National Library of Medicine has published research on the accuracy of consumer-grade accelerometers and wearables that consistently finds step detection to be reasonably accurate but step-to-distance conversion to be the primary source of error in derived metrics.
The practical takeaway
Step counts are reliable as counts. They are unreliable as implicit measures of distance, calorie burn, or activity volume, because they hide the stride length variable that determines what a step actually means in physical terms.
For fitness goals stated in distance, mileage, or calorie units -- which is most serious fitness goals -- converting to a personalized distance figure using your actual height and activity type produces substantially more accurate data. The conversion is a one-time setup that makes all subsequent tracking more meaningful.
For the full guide on using the conversion to set distance-based goals, see How to Convert Your Step Count Into Real Distance Goals.
How Devices Try to Correct the Error
Modern wearables and health platforms have introduced several approaches to reduce the stride length error:
Height input. Most devices allow you to enter your height in a profile settings screen. This enables the device to apply a height-specific stride estimate from the formula discussed above. The improvement is meaningful but not complete -- the formula is still an approximation, and activity-type variation is not fully addressed by a static height input.
GPS calibration. When you take a GPS-tracked outdoor walk or run, devices like Apple Watch and Garmin watches can compare their step count to the GPS-measured distance and adjust their stride estimate accordingly. This is the most accurate calibration method and produces substantially better distance estimates for the specific activity type being calibrated. The calibration typically applies only to the activity type used for calibration (walking vs. running), and GPS accuracy varies based on satellite coverage.
Machine learning activity classification. Newer wearables use accelerometer data patterns to classify activity type (walking, running, cycling, stair climbing) and apply different stride models per activity type. The classification accuracy depends on training data quality and the distinctiveness of the accelerometer signal for each activity. Transitions between activity types remain the weakest point in the classification.
Why the Error Matters More for Some Users Than Others
The stride length error is most significant for three groups:
People at height extremes (under 5 feet or over 6 feet 2). For these users, the population-average assumption differs from their actual stride by 15 to 20 percent, and every derived metric inherits that error.
Active people who mix walking and running regularly. If a device's activity type classifier misidentifies a run as a walk (or vice versa), the stride estimate is wrong by 30 to 45 percent for that activity, and the distance and calorie estimates are substantially off.
People using step-based data for medical or nutritional management. For users tracking activity as part of a diabetes management program, cardiac rehabilitation, or weight management protocol, the accuracy of derived calorie estimates matters more than it does for casual fitness tracking.
When to Use GPS vs. Step-Based Distance
For outdoor activities where GPS is available and accurate, GPS-based distance is always more accurate than step-count-derived distance. GPS measures distance directly from position changes; step counting infers distance from motion patterns.
For indoor activities, gym workouts, or situations where GPS is unavailable (subways, large buildings), step counting is the only available option. In these cases, a personalized conversion factor -- from the EvvyTools Steps to Distance Calculator or from your device's height-specific formula -- is the best available approximation.
For daily step counting across mixed indoor and outdoor activity, a hybrid approach works well: use GPS distance for outdoor activities where you actively start a workout, and use a personalized step-to-distance ratio for background step counts throughout the day.
For the full guide on converting step counts to meaningful distance goals, see How to Convert Your Step Count Into Real Distance Goals.
How Devices Try to Correct the Error
Modern wearables and health platforms have introduced several approaches to reduce the stride length error:
Height input. Most devices allow you to enter your height in a profile settings screen. This enables the device to apply a height-specific stride estimate from the formula discussed above. The improvement is meaningful but not complete -- the formula is still an approximation, and activity-type variation is not fully addressed by a static height input.
GPS calibration. When you take a GPS-tracked outdoor walk or run, devices like Apple Watch and Garmin watches can compare their step count to the GPS-measured distance and adjust their stride estimate accordingly. This is the most accurate calibration method and produces substantially better distance estimates for the specific activity type being calibrated. The calibration typically applies only to the activity type used for calibration (walking vs. running), and GPS accuracy varies based on satellite coverage.
Machine learning activity classification. Newer wearables use accelerometer data patterns to classify activity type (walking, running, cycling, stair climbing) and apply different stride models per activity type. The classification accuracy depends on training data quality and the distinctiveness of the accelerometer signal for each activity. Transitions between activity types remain the weakest point in the classification.
Why the Error Matters More for Some Users Than Others
The stride length error is most significant for three groups:
People at height extremes (under 5 feet or over 6 feet 2). For these users, the population-average assumption differs from their actual stride by 15 to 20 percent, and every derived metric inherits that error.
Active people who mix walking and running regularly. If a device's activity type classifier misidentifies a run as a walk (or vice versa), the stride estimate is wrong by 30 to 45 percent for that activity, and the distance and calorie estimates are substantially off.
People using step-based data for medical or nutritional management. For users tracking activity as part of a diabetes management program, cardiac rehabilitation, or weight management protocol, the accuracy of derived calorie estimates matters more than it does for casual fitness tracking.
When to Use GPS vs. Step-Based Distance
For outdoor activities where GPS is available and accurate, GPS-based distance is always more accurate than step-count-derived distance. GPS measures distance directly from position changes; step counting infers distance from motion patterns.
For indoor activities, gym workouts, or situations where GPS is unavailable (subways, large buildings), step counting is the only available option. In these cases, a personalized conversion factor -- from the EvvyTools Steps to Distance Calculator or from your device's height-specific formula -- is the best available approximation.
For daily step counting across mixed indoor and outdoor activity, a hybrid approach works well: use GPS distance for outdoor activities where you actively start a workout, and use a personalized step-to-distance ratio for background step counts throughout the day.
For the full guide on converting step counts to meaningful distance goals, see How to Convert Your Step Count Into Real Distance Goals.
Additional reading: the World Health Organization physical activity fact sheet covers the global evidence base for aerobic activity recommendations, and the National Institutes of Health physical activity research portal provides access to current exercise science studies.
Top comments (0)