Health Tracking on Your Wrist: How Accurate Is It, Really?

Consumer smartwatches use PPG sensors — tiny LEDs that detect blood flow changes through the skin. At rest, published research has generally found error rates within 5% compared to ECG references. During intense exercise, however, wrist movement creates signal interference (called motion artifact), and error margins can widen significantly. Chest strap monitors, which measure electrical heart signals directly, remain more reliable during dynamic activity for users who need precision.

Clinically validated sleep analysis uses PSG, which records brain waves via electrodes on the scalp. Wearables have no access to brain activity; they infer sleep stages from accelerometer data and heart rate patterns. Studies comparing consumer devices against PSG have found that wearables are reasonably good at distinguishing sleep from wakefulness, but less reliable when differentiating specific stages like N3 (deep sleep) from lighter stages. Treat sleep stage breakdowns as useful directional data, not precise measurements.

Medical pulse oximeters are FDA-cleared devices calibrated across a range of oxygen saturation levels, including low-oxygen conditions where precision is most critical. Most consumer wearables use SpO2 measurement primarily for sleep monitoring features and are not intended or cleared for medical diagnosis. Several independent evaluations have noted that wrist-based SpO2 readings can deviate by several percentage points, particularly on users with darker skin tones — a known limitation tied to how optical sensors interact with melanin. If you have a medical reason to monitor blood oxygen, consult a healthcare provider about appropriate tools.

Calorie calculations on wearables rely on algorithms that combine heart rate, movement data, and user-entered profile information (age, weight, height, sex). Each input introduces its own uncertainty, and the algorithms themselves are trained on population averages that may not reflect individual metabolic variation. Activities with irregular movement patterns — cycling, weightlifting, rowing — tend to produce larger errors than steady-state walking or running. Treating calorie data as a rough relative indicator rather than an absolute number reduces the risk of over- or under-correcting your diet based on flawed output.

HRV — the variation in time between consecutive heartbeats — is a legitimate area of physiological research linked to autonomic nervous system function. However, consumer wearables measure HRV using optical sensors (less precise than ECG), apply proprietary algorithms to generate a single "stress" or "recovery" score, and rarely disclose how those algorithms work. Different devices can produce markedly different scores from the same underlying signal. These metrics can surface useful patterns over time, but comparing your score to a clinical reference range or another person's device output is not meaningful.