During the keynote launch of the Apple Watch Series 12 and Ultra 4, Apple claimed both watches offer the most accurate heart rate sensing of any wearable. The backing for that line is an Apple-conducted study with more than a thousand people, a Polar H10 chest strap as the reference, and a head-to-head against six other current wearables, including the Garmin Forerunner 970.

Apple’s paper puts a tighter frame on the same claim. It says Apple Watch “offers users the most accurate heart rate sensing among the wearable devices evaluated in this study.”

How Apple ran it

Apple designed the study and approved the consent form in-house. Testing ran in July and August 2026 at five sites: two near Cupertino, plus San Diego, Austin, and Selangor, Malaysia. Apple ran Cupertino and Austin. A contract research group ran the other two, and 59% of the people in the final analysis came from those sites.

Apple enrolled 1,460 people and used 1,254 who produced at least one paired comparison. The group was 35% female, with a mean age of 39 and a mix of body types and skin tones, including 20% with darker Fitzpatrick skin tones (V or VI). Participants had to be 18 or older and able to wear a watch, a ring, and a chest strap. Apple left out tech-media workers, people on beta blockers, anyone who shouldn’t do intense exercise, and anyone with a tattoo or large mole under a sensor.

What they wore

Each person wore an Apple Watch on one wrist, another device on the other, and a Polar H10 chest strap. Wrist side was randomized, so the Apple Watch sat on the dominant hand about half the time. Staff fitted every device to the manufacturer’s recommendation and used the latest retail model from each brand.

The lineup was Apple Watch Series 12, Garmin Forerunner 970, Google Pixel Watch 4, Huawei Watch 5, Samsung Galaxy Watch 8, WHOOP 5, and Oura Ring 5 on a finger. Apple says those watch brands cover about 90% of worldwide smartwatch shipments, plus two screenless devices. Oura sat out strength training because Oura tells people not to lift in the ring.

Workouts ran 15 to 20 minutes: outdoor runs, indoor interval treadmill, outdoor cycling, HIIT, and strength. Apple picked those because they are hard on wrist heart rate — bouncing cadence, a loaded grip, fast arm motion. Everyday sessions were about 20 minutes of rest, desk work, meal prep, and walking. Workout mode stayed on for workouts and off for the everyday stuff.

How Apple scored it

Every reading was compared with the Polar strap. Apple used two error scores, RMSE and MAE, plus the average heart rate for the whole session. In plain terms, those first two numbers measure how far a device drifted from the chest strap, with RMSE punishing a few big misses more. Apple then compared each person’s Apple Watch error with the other device on their other wrist, and it treated a result as significant only if a 99% confidence interval stayed on Apple’s side. At least 50 people were required for each workout and each competitor.

Apple says it wanted the heart rate a user actually sees, at the highest rate the product offers. A lot of the other devices were logged over Bluetooth onto an iPhone. Apple Watch background heart rate only lands in HealthKit every 30 seconds, so Apple used an internal logger that grabbed the same 5-second number shown on the watch face and in the Workout app. Samsung’s everyday export also skipped a proper second-by-second stream, so Apple compared one exported number with the chest-strap average for that stretch.

What Apple reported

On the overall workout and everyday scores, Apple says Series 12 was more accurate than every other device on both error measures. Session-average heart rate went Apple’s way in 11 of 12 of those rollups. The exception is Huawei Watch 5 on overall workouts, where Apple Watch looked a bit better but the gap was too small to call.

Broken out by workout, Apple says Series 12 won every activity against every device except outdoor cycling versus Pixel Watch 4, which was too close to call. Everyday use looked similar. Eight rest, desk, and walking comparisons were too close to call, and Pixel Watch 4 beat Apple Watch at rest only, by less than half a beat.

Added up, Apple counted 159 individual comparisons. Series 12 was clearly better in 139, 18 were too close to call, and the other device won two. If you ignore the confidence cutoff and just look at the raw direction, Apple says it came out ahead in 155 of 159.

Study limitations

Apple designed the study, approved the consent form, and ran two of the five sites. The published charts show the gap versus Apple Watch, not a simple “this watch was off by X bpm” number you can hang on a Forerunner 970 or anything else in the lineup. Devices were also logged in different ways, and the keynote claim only covers the wearables and activities Apple chose to test.

Taken on its own terms, the paper matches what Apple said on stage: in this head-to-head, Series 12 came out ahead of the other devices it tested.

Source: Apple Watch Heart Rate Accuracy Study, September 2026. Apple also summarizes the claim in its health newsroom post.