A fitness tracker measures three things directly and infers everything else. Knowing which numbers are measured and which are estimated changes how much weight to put on them.
What is actually measured
Movement. An accelerometer detects acceleration in three axes, and usually a gyroscope detects rotation. This is genuinely measured, and step counting from it is reliable to within a few per cent for walking.
Blood flow at the skin. An optical sensor shines green LEDs into the wrist and measures how much light comes back. Blood absorbs green light, so the reflected signal pulses with each heartbeat. This is the basis of wrist heart rate, and it is a real measurement — of blood volume at the skin, from which heart rate is derived.
Location, if it has GPS. Distance and pace outdoors come from satellites and are accurate to a few metres.
Everything else on the screen is calculated from those three, plus your height, weight, age and sex.
What is estimated, and how well
Calories burned. This is a model, not a measurement. It combines heart rate, movement and your stated body data against population averages. Research comparing consumer trackers against laboratory calorimetry has repeatedly found errors of 20–30 per cent, sometimes more, and the error is not consistent between people.
Use the trend, not the number. If it says 400 today and 600 tomorrow, you did more. If it says 500, that is not 500.
Sleep stages. Light, deep and REM sleep are defined by brain activity, which a wrist device cannot see. Trackers infer stages from movement and heart rate variability. Total sleep time and time in bed are reasonably accurate; the breakdown into stages is much weaker, and agreement with clinical sleep studies on stage classification is often poor.
Sleep duration and consistency are the useful outputs here. The hypnogram is entertainment.
Stress and recovery scores. Derived from heart rate variability, which is a real physiological signal with genuine meaning — but the mapping from HRV to a single "readiness" number is a proprietary model, different on every brand, and not comparable between them.
VO2 max. Estimated from the relationship between your pace and your heart rate. Directionally useful over months, not a lab value.
Where wrist heart rate struggles
Optical wrist sensing has known weak points, and they are all about the sensor losing sight of a clean pulse:
- High-intensity intervals. Heart rate changes faster than the algorithm smooths, so readings lag at the start of a hard effort.
- Cold weather. Reduced blood flow to the skin weakens the signal.
- Loose fit, or wearing it over the wrist bone. The single most common cause of bad data. A tracker should sit a finger's width above the wrist bone, snug enough not to slide.
- Gripping activity. Rowing, cycling on the drops, weight training — flexing the forearm moves the sensor against the skin.
- Tattoos and darker skin tones can both reduce reflected green light, though modern multi-wavelength sensors have narrowed this.
A chest strap measures the heart's electrical signal directly and is materially more accurate for intervals and lifting. If heart rate is the number you actually train by, a strap is the upgrade that matters, not a more expensive watch.
Step counts and the 10,000 myth
Step counting is one of the reliable outputs. The 10,000 figure, however, came from a 1960s Japanese pedometer marketing campaign, not from research.
Later studies have found meaningful health benefits plateauing somewhere between 7,000 and 8,000 steps a day for most adults. 10,000 is a fine target. It is not a threshold, and falling short of it is not a failure.
What a tracker is genuinely good for
Consistency, trends and prompts. It is reliable at telling you that this week had less movement than last week, that you slept ninety minutes less than usual, and that you have been sitting for two hours.
Those are behavioural signals and they are what the device is actually good at. Treating the calorie figure as a budget, or the sleep score as a diagnosis, is asking it for a precision it does not have.
None of this is medical advice, and no consumer tracker is a diagnostic device.
More training notes are filed under Health & Fitness.