Last updated July 2026.
You finish a run. The treadmill says 420 calories. Your Apple Watch says 310. Your Fitbit, if you’re wearing both like a crazy person for science, says something else entirely. Nobody’s lying to you, but nobody’s exactly right either. Here’s why these numbers almost never match, and which one deserves more of your trust.
Two Completely Different Measurement Methods
A treadmill console estimates your calorie burn from what it can directly control and measure: your speed, your incline, and your time, usually combined with a default or manually entered body weight. It’s a physics-based estimate, essentially a version of the MET formula explained on our methodology page.
A wrist-worn device like an Apple Watch or Fitbit works completely differently. It estimates calorie burn primarily from your heart rate, using an optical sensor to detect blood flow changes, then runs that heart rate data through the device’s own proprietary algorithm. It’s an indirect, physiology-based estimate, not a movement-based one.
Two different inputs, two different formulas, run by two different companies. Disagreement isn’t a malfunction, it’s the expected outcome.
What the Research Actually Found
In 2017, researchers at Stanford University School of Medicine ran the most cited independent test of wearable device accuracy to date. They evaluated seven popular devices, including the Apple Watch, Fitbit Surge, Microsoft Band, and others, against laboratory-grade measurement in 60 volunteers who walked, ran, and cycled.
The results split sharply by metric:
| Metric | Finding |
|---|---|
| Heart rate accuracy | 6 of 7 devices measured within about 5% error, better than researchers expected |
| Calorie/energy expenditure accuracy | Even the best-performing device (Fitbit Surge) was off by an average of 27%. The worst-performing device was off by 93% |
| Researchers’ accuracy bar | Under 10% error was considered acceptable for lay use; none of the seven devices hit that bar for calories |
The takeaway from the study’s senior author, Dr. Euan Ashley: these devices are good at measuring heart rate and not good at converting that heart rate into an accurate calorie number. That gap has narrowed somewhat since 2017 as algorithms have improved, but the fundamental limitation, estimating energy expenditure from heart rate alone, hasn’t gone away.
So Which Number Should You Trust?
Neither one is a lab-verified fact, so “trust” isn’t really the right frame. What matters more is picking one method and sticking with it for tracking trends over time, rather than jumping between your console and your watch and getting confused by the gap. If you want a number grounded in a documented, checkable formula rather than a proprietary heart rate algorithm, a MET-based calculator (like the one on our homepage) at least shows you exactly how the number was produced.
Related Reading
- Why treadmills overestimate calories burned
- Heart rate vs. MET-based calculators, compared
- Our full calculation methodology
Sources
- Shcherbina, A. et al. (2017). Accuracy in Wrist-Worn, Sensor-Based Measurements of Heart Rate and Energy Expenditure in a Diverse Cohort. Journal of Personalized Medicine, published via Stanford University School of Medicine.