Are fitness tracker step counts accurate?

Fitness trackers are useful tools, but the step count on your wrist is not a laboratory measurement. How accurate it is depends almost entirely on how fast you walk and what your hands are doing.

The accuracy picture at different speeds

Speed is the single biggest predictor of wrist-tracker accuracy. At a brisk pace, most mainstream devices do well. At a slow shuffle, the error climbs fast.

A 2024 meta-analysis put slow-walking mean absolute percent error at 23.9% versus 3.0% for brisk walking. That gap is not a firmware bug. It is a physics problem: at slow speeds, each step produces a smaller wrist-acceleration signal, and the device has a harder time separating steps from general arm movement.

Why pushing a cart breaks wrist trackers

The wrist accelerometer logs steps by detecting the rhythmic acceleration of arm swing. When your hands grip a shopping cart, stroller, or wheelchair, that swing stops. Your legs keep moving. Your wrist does not.

Research comparing wrist and hip monitor performance found that wrist-worn trackers underestimate steps during cart-pushing by 35% to 95%. That is not a rounding error. A 10,000-step shopping trip could log anywhere from 500 to 6,500 steps on your watch.

The same problem appears with assistive devices. A study on Fitbit accuracy with walkers and rollators found wrist-position error of 31.2% on average compared to 1.5% error for hip and ankle positions with the same device.

The practical fix: for any activity where your hands grip something stationary, clip the tracker to your waistband or slide it into a pants pocket. That one change eliminates most of the error.

Hip placement is the gold standard

The research on placement is consistent across studies. A comparison of wrist-worn and hip-worn Actigraph GT3X devices in free-living conditions found the wrist version recorded approximately 39% more steps per day than the hip version (11,203 vs 6,866 steps on average), with the gap widening in older adults to nearly a 2:1 ratio.

That 39% overcounting at the wrist happens because the wrist picks up non-step arm movements and logs them as steps. A hip clip sits close to your center of mass and measures the actual vertical and horizontal movement of your body with far less noise.

The wrist picks up arm movements that are not steps and misses steps when the arm stays still. Hip placement eliminates most of that noise with no change to your device.

For people who walk slowly, use assistive devices, or push a cart regularly, a dedicated hip-clip pedometer is worth considering alongside a wrist tracker. The wrist device wins on convenience and features; the hip clip wins on step-count accuracy.

What "accurate enough" actually means

At normal walking speeds with free arm swing, the error on a mainstream wrist tracker like the Fitbit Inspire 3 is small enough that your goals and trends remain meaningful. A 3% to 10% error at a 10,000-step target means your tracker might read 9,000 to 11,000 real steps. That margin does not change whether your trend is up or down week over week.

The 10,000-step target stays useful for motivation and relative comparison even if the absolute count is off. What matters is whether the number is consistently off in the same direction. A tracker that always undercounts by 8% still shows whether you walked more this week than last.

Accuracy matters more when you are recovering from injury, monitoring activity for a health condition, or comparing your numbers directly against a clinical study's step targets. In those cases, hip placement and calibration are worth the effort.

How to calibrate for better results

FAQ

If you are choosing a device for walking fitness, see our guide to the best fitness trackers for walking for picks across price points and use cases. Browse more fitness gear or read how we research and rate.