A smartwatch shows very precise numbers: 8,432 steps, 64 beats per minute, 6 hours and 47 minutes of sleep. The apparent precision of the digits, however, is not the same as the accuracy of the measurement. Behind every value there is a sensor detecting something indirect and an algorithm interpreting it. Knowing how they work helps you use them well and not trust them blindly.
The sensors inside the strap
Accelerometer and gyroscope
The accelerometer measures acceleration along three axes; the gyroscope measures rotation. Together they record every movement of the wrist. They are the basis for counting steps, recognising a workout, detecting a fall or telling whether you are lying still in bed.
Optical heart rate sensor (PPG)
On the back of the watch there are LEDs, usually green, and photodiodes. This is photoplethysmography: light enters the skin and part of it is absorbed by the blood. With every heartbeat the volume of blood in the capillaries changes slightly, and so does the amount of reflected light. The device turns these small fluctuations into a heart rate.
Some models also use red and infrared LEDs to estimate oxygen saturation, exploiting the different way oxygenated and deoxygenated haemoglobin absorb light.
Other sensors
- GPS or GNSS, for distance and speed when running and cycling;
- barometer, for floors climbed and elevation gain;
- skin temperature sensor, used to observe changes relative to your own baseline;
- electrodes, on some models, to record a single-lead electrocardiogram when you rest a finger on the crown.
Where the numbers come from
Steps
The algorithm looks in the accelerometer signal for a repetitive pattern consistent with walking. That is why pushing a trolley or a pram, with the wrist still, can lose steps, while gesturing or driving on bumpy roads can add some. Over a whole day the errors partly cancel out.
Heart rate
At rest the optical sensor is generally reliable. During intense activity with sudden wrist movements, such as weightlifting, tennis or combat sports, motion disturbs the reading and values can be off. A chest strap, which measures the heart's electrical activity, remains more accurate under heavy effort.
Practical factors matter too:
- a strap that is too loose lets outside light in;
- wrist tattoos can block the passage of light;
- cold reduces peripheral circulation and weakens the signal.
Sleep
The watch does not measure sleep in the strict sense: it infers it from the absence of movement, heart rate and heart rate variability. The "light, deep, REM" stages are estimates derived from these signals, not a recording of brain activity. Detecting when you fall asleep and wake up works fairly well; the breakdown into stages should be taken with caution, especially when comparing devices from different brands.
Reading the data usefully
The value of a wearable lies not in a single number but in the trend. A concrete example: if your resting heart rate is stable for weeks around a certain value and then rises by several beats for a few days, that change says something about fitness, stress or a run of poor sleep.
A few practical rules:
- compare your data with yourself, not with other people;
- look at weekly averages rather than daily values;
- always wear the watch the same way, a finger's width from the wrist bone;
- do not switch devices and expect the same numbers: different algorithms give different results.
Battery and privacy
GPS and continuous heart rate and oxygen measurement are the most power-hungry features. Turning off continuous saturation tracking and limiting the always-on display can double battery life on many models.
The data collected is personal and reveals a lot about your habits: sleep times, workout locations, physical condition. It is worth checking in the app settings who can see your activities, turning off public route sharing and reviewing which third-party apps have access to the data.
In short
- Accelerometer, gyroscope and the PPG optical sensor are the heart of every wearable.
- Optical heart rate is reliable at rest, less so during sudden wrist movements.
- Sleep stages are estimates derived from movement and heart rate, not direct measurements.
- The data is most useful for observing personal trends over time.
- GPS and continuous measurements are the features that drain the battery most.
