Smart rings versus watches: your fingertip and your wrist are not telling the same story

Smart rings have gone mainstream, and the question I now get at dinner tables is always the same: ring or watch, which is more accurate? The honest answer is more interesting than either camp admits, because where a sensor sits genuinely changes what it can see.

This year a study in the journal Sensors measured optical signal quality across the body and found the fingertip the best site of all, and the wrist the most challenging of the lot (Sensors, 2026). That is the physics behind the whole debate, finally measured directly.

Rings and watches both work by shining light into your skin and reading the tiny changes as blood pulses through. The fingertip is one of the most richly supplied places on your body, which is why the clip a hospital uses to read your pulse goes on your finger, not your wrist. The wrist is the hardest site of the lot: in that same study it came last, winning on none of the quality measures (Sensors, 2026).

The published record matches the physics, with caveats worth knowing. Overnight, the Oura ring's readings tracked a reference monitor very closely across a whole night, though agreement loosened considerably when the same nights were cut into five-minute windows (Kinnunen 2020, by Oura-affiliated scientists; partly replicated by an independent group in 2022). During exercise, wrist devices wobble: in lab testing, a chest strap stayed true while wrist readings drifted, and no wrist device stayed accurate on an elliptical with moving arm levers (Gillinov 2017).

Now the honesty section. The head-to-head everyone assumes exists barely does. The main published ring-versus-watch comparison on the same people covers sleep only, and it was funded by Oura, whose medical advisory board the lead author sits on (Robbins 2024: the ring edged the watches). I am not aware of a comparable study for exercise. And rings have their own physics problem: cold. Chilly fingers pull blood away from the surface, and in cold-exposure testing the finger signal fell by more than 80%. Watches, meanwhile, struggle with tattooed wrists; Apple says itself that tattoo ink can block the sensor.

And there is a bigger limitation than either of those, which this field was slow to talk about. Green light, which most wrist devices use, is absorbed by melanin, so darker skin returns a weaker signal. In one study the same kind of wrist sensor was out by about one beat per minute on the lightest skin and about eleven on the darkest (reviewed in Fine 2021). The evidence is not perfectly consistent, and some devices using red and infrared light show no clear difference, but the direction is established well enough that a 2020 editorial in SLEEP called it an underreported problem with real consequences for who these devices actually work for (Colvonen 2020).

Match the device to the job. For overnight patterns, resting rhythms, temperature trends, recovery, the finger is a naturally strong site. For exercise, neither is a chest strap; a snug watch does well for steady, rhythmic movement and worse for anything with erratic arm action. If your hands run permanently cold, a ring will struggle; if you have a wrist tattoo under the sensor, a watch will.

Two rules hold whichever you wear. Trends beat absolutes: a device that is consistently a little off is still excellent at spotting change, and change from your own baseline is where the insight lives. And your wearable reads patterns, not causes: if your heart is doing something that worries you, racing at rest, thudding, making you feel faint, that is a conversation for your doctor, not a graph.