Smartwatch HR Accuracy During Strength Training

Target audience: This article is for people who lift weights while wearing a smartwatch, fitness tracker, or sports watch. It is also for coaches, gym beginners, strength athletes, health writers, and anyone who has stared at a wrist display after a hard set and thought, “That number cannot be right.” No technical background is needed. The goal is to explain smartwatch heart rate accuracy during strength training in plain language, without treating the watch like either a miracle gadget or a useless toy.

 

Key points covered: wrist-based heart rate sensors can be useful during lifting, but they have limits. Most smartwatches estimate heart rate with photoplethysmography, or PPG. That means the device shines light into the skin and reads changes in blood volume near the wrist. The method is convenient, but lifting is a messy environment for clean data. Grip pressure, wrist flexion, sweat, skin movement, short sets, rest periods, and fast changes in effort can all interfere with the signal. Research shows that heart rate is usually one of the stronger smartwatch metrics, especially when compared with calorie estimates. It also shows that accuracy depends on the device, the activity, the person, and the question being asked. A watch may be good enough for tracking trends across sessions. It may still miss a short heart rate spike during a heavy set.

 

The first thing to understand is that a smartwatch is not measuring the heart in the same way as an electrocardiogram. An ECG records the electrical activity that coordinates each heartbeat. A wrist sensor usually reads blood-volume changes under the skin. That difference matters. It is the gap between hearing the drummer directly and trying to follow the beat through a wall while someone is moving furniture. During quiet sitting, the wall is not a big problem. During squats, rows, pull-ups, kettlebell swings, and loaded carries, the wall starts shaking.

 

A 2023 review by Kim and Baek, “Photoplethysmography in Wearable Devices: A Comprehensive Review of Technological Advances, Current Challenges, and Future Directions,” explains the basic technology. The authors describe PPG as a noninvasive optical method that uses a light source and a photodetector to detect changes in blood volume in tissue. In wrist devices, the signal comes from reflected light rather than a direct electrical reading from the heart.1 That is why contact quality matters. If the sensor shifts, presses too hard, floats loosely, or sits over moving tissue, the device has to separate the pulse signal from noise.

 

Strength training makes that separation difficult because the wrist is rarely relaxed. During a bench press, the wrist may extend under load. During a deadlift, the hand crushes the bar. During curls, rows, and pull-ups, the forearm muscles contract repeatedly. Those contractions can change local blood flow and tissue pressure near the sensor. The watch is trying to read a pulse wave while the wrist is acting like a busy construction site. The signal may still be usable, but it is not the same environment as steady cycling on an indoor bike.

 

Running and lifting also create different heart rate patterns. A steady run often produces a gradual climb in heart rate, followed by a fairly stable rhythm. Strength training is more like a city traffic light. Effort jumps during the set. Breathing changes. The lifter may brace the trunk. The set ends. Heart rate may continue rising for a short time, then fall during rest. The watch algorithm may smooth those changes to avoid reporting nonsense every second. That smoothing can make the displayed value look late, flat, or lower than expected.

 

The research base supports that practical observation. In the 2020 npj Digital Medicine study “Investigating Sources of Inaccuracy in Wearable Optical Heart Rate Sensors,” Bent, Goldstein, Kibbe, and Dunn tested consumer and research-grade wearable optical heart rate sensors under multiple conditions. The study found no statistically significant difference in accuracy across skin tones, but it did find significant differences between devices and activity types. The authors reported that absolute error during activity was about 30% higher than during rest.2 This does not mean wrist devices fail during exercise. It means exercise changes the measurement problem.

 

That same study is useful because it avoids a lazy answer. The issue is not simply “dark skin makes watches inaccurate” or “one brand is always right.” The authors identified several sources of PPG inaccuracy, including motion artifacts, signal crossover, and device-specific behavior.2 Motion artifact is the key phrase for lifters. It means movement adds patterns to the signal that are not caused by heartbeats. During a strict machine exercise, motion artifact may be lower. During burpees with dumbbells, the sensor is riding a mechanical bull.

 

A more direct strength-training study appeared in 2026 in Sensors. Lee, Jun, Bae, Roh, and Cho published “Comparative Validity of Smartwatch-Derived Heart Rate and Energy Expenditure During Endurance and Resistance Exercise.” The study included 62 healthy adult men. Participants performed standardized endurance and resistance exercise protocols while wearing Apple, Galaxy, Fitbit, and Garmin smartwatches. Heart rate was compared with ECG. Energy expenditure was compared with indirect calorimetry, which estimates metabolic cost from respiratory gas exchange.3

 

The results were not a simple brand-war headline. During endurance exercise, heart rate measurements from all smartwatch brands were comparable with ECG. During resistance exercise, only the Apple Watch showed no significant difference from ECG in that study. Across devices, heart rate readings showed strong correlations with ECG, with r values from 0.64 to 0.97. Reliability was reported as excellent, with intraclass correlation coefficients above 0.94. Limits of agreement were about plus or minus 10 beats per minute.3 That sounds useful, but it also means the exact number during one short set should not be treated like lab truth.

 

The same 2026 study gives a sharper warning about calories. Energy expenditure accuracy was limited across all devices. During resistance exercise, the correlations between smartwatch energy expenditure and indirect calorimetry were weak, with r values from 0.10 to 0.34. Reliability was poor, with intraclass correlation coefficients below 0.45.3 In plain English, the heart rate number may be decent for tracking effort, but the calorie number during lifting can wander into fantasy territory. A watch may tell you a leg day burned enough energy to justify a banquet, or it may act as if you only rearranged pencils. Neither reading deserves blind trust.

 

There is a reason calorie estimates struggle. Resistance training is not just a heart-rate event. It includes mechanical work, muscular tension, pauses, breathing changes, rest intervals, and sometimes a Valsalva maneuver, where the lifter holds breath and braces during effort. Two people can show similar average heart rates during a session while doing different amounts of external work. One may perform heavy squats with long rest. The other may do circuit training with light loads and short rest. A smartwatch sees motion, pulse, and user profile data. It does not truly know bar speed, load accuracy, range of motion, tendon stress, or how close the set was to failure.

 

Systematic reviews point in the same direction. Fuller and colleagues reviewed commercially available wearable devices for steps, energy expenditure, and heart rate in JMIR mHealth and uHealth in 2020. The review found that heart rate was generally more reliable than energy expenditure, while calorie estimates varied across devices and activity contexts.4 Zhang and colleagues published a 2020 systematic review and meta-analysis in the Journal of Sports Sciences on wrist-worn PPG devices for heart rate. It included 44 articles and 738 effect sizes across 15 brands, comparing wrist estimates with criterion measures such as ECG or chest straps.5 The overall message is consistent: heart rate is one of the more defensible wearable metrics, but the wrist location and the exercise mode matter.

 

A 2025 study by Van Oost and colleagues, “Accuracy of Heart Rate Measurement Under Transient States: A Validation Study of Wearables for Real-Life Monitoring,” adds another piece. The researchers compared one chest-worn device, one research-grade wrist device, and five commercial wrist devices with a 12-lead ECG during a 20-minute protocol designed to simulate real-life changes, including rest and walking at varied intensity. The devices included Zephyr BioHarness 3.0, EmbracePlus, Fitbit Charge 5, Fitbit Sense 2, Garmin Vivosmart 4, WHOOP 4.0, and Withings ScanWatch.6 The practical takeaway is not that walking equals lifting. It is that rapid transitions expose device behavior. A device can look acceptable when data are averaged but lag or drift when the body changes pace.

 

That distinction matters in the gym. Average heart rate for a 50-minute session may be close enough to show whether the workout was calm, moderate, or stressful. Set-by-set data may be less dependable. A watch can miss the peak of a heavy triple because the effort is short. It may display the peak after the set ends because the algorithm catches up late. It may also show a sudden drop if the wrist bends, the strap shifts, or sweat changes skin contact. That strange number is not always a sign that your heart behaved strangely. It may be a sign that the device lost a clean optical signal.

 

Chest straps usually perform better for rapid exercise heart rate because they measure electrical signals closer to the heart. They are not flawless. Poor electrode contact, dry skin, strap movement, and battery issues can still cause errors. But for lifting intervals, mixed conditioning, or heart-rate-zone training, a chest strap often has an advantage over the wrist. ECG remains the reference method in validation studies, not because gym users need medical equipment, but because ECG gives researchers a clearer comparison point.

 

Device brand still matters, but it should be treated carefully. Apple Watch, Samsung Galaxy Watch, Fitbit, Garmin, WHOOP, and Withings devices use different sensors, firmware, sampling choices, and algorithms. They also change over time. A study of one model does not automatically apply to every older or newer version. A software update can alter signal processing. A case size can affect fit. A tattoo, wrist shape, strap material, and placement can change contact. This is why a ranking list from one test may not survive contact with your next workout. The more useful question is narrower: does this device give repeatable information for this person, during this type of training, under these wearing conditions?

 

The critical perspective is simple. Wearable fitness data can help, but it can also create false precision. A smartwatch number looks clean because it appears as digits on a screen. Digits feel official. That is the trap. A display reading of 137 beats per minute looks more serious than “somewhere around moderate effort,” but the real uncertainty may be larger than the screen suggests. During strength training, the watch is not a judge with a gavel. It is a witness standing near the back of the room, trying to hear through noise.

 

This matters because people change behavior based on wearable data. Some lifters cut rest too short because the watch says their heart rate has dropped. Others feel disappointed when a hard session receives a low calorie count. A beginner may think the workout “didn’t count” because the heart rate graph looks unimpressive. That reaction is understandable. Modern apps turn training into scores, rings, badges, and charts. The gym starts to feel like a video game, except the controller is strapped to your wrist and sometimes forgets you just fought a barbell dragon.

 

The emotional part deserves attention because data can mess with motivation. A hard set of squats can feel like a tax audit conducted inside your lungs. If the watch shows a low number afterward, the mismatch can be annoying. The correct response is not to throw the watch into a locker like a tiny traitor. The better response is to understand what it can and cannot see. A low reading during a set may reflect sensor limitations. It does not erase mechanical tension, muscle recruitment, joint loading, or training volume.

 

To get better smartwatch heart rate readings during strength training, start with fit. Wear the device snugly, but not so tight that it causes discomfort or compresses tissue heavily. Place it slightly above the wrist bone, where the case can sit flat against the skin. Warm up before judging readings, because cold skin and low peripheral blood flow can make optical signals harder to read. Keep the sensor clean. Tighten the strap before training, then loosen it afterward if needed. If the exercise bends the wrist hard, expect more noise. If the movement requires heavy gripping, expect another challenge.

 

Next, change what you pay attention to. Do not obsess over one-second readings during reps. Look at session averages, recovery patterns, and repeated trends across similar workouts. Compare Monday squats with future Monday squats, not with a long run or a yoga class. If your watch allows manual workout selection, choose strength training rather than letting the device guess. If accurate heart rate zones are important for conditioning circuits, sled pushes, rowing intervals, or hybrid training, use a compatible chest strap when available. For calorie estimates, use the number as a rough app output, not as a food budget.

 

A practical checking method is to compare devices against your body signals in a structured way. During a warm-up, note whether the watch rises gradually as effort increases. During rest periods, check whether it falls in a predictable direction. During repeated sets with similar loads and rest, look for a consistent pattern. If readings jump from 90 to 170 and back to 95 with no matching change in effort, the signal is likely unstable. If the heart rate stays flat during high-effort conditioning, check strap position and sensor contact. If symptoms appear, stop using the watch as the deciding authority.

 

Symptoms outrank wrist data. Chest discomfort, pressure, fainting, unusual shortness of breath, severe dizziness, or palpitations that feel abnormal should not be dismissed because the watch shows a normal number. The American Heart Association lists chest discomfort, shortness of breath, discomfort in other upper-body areas, nausea, lightheadedness, and unusual fatigue among warning signs that require attention.7 This article is about measurement limits, not diagnosis. A consumer wearable can miss, delay, or misclassify signals. Your body gets a vote, and in urgent situations it gets the louder one.

 

The best role for smartwatch heart rate during lifting is trend tracking. It can show whether your conditioning work is becoming easier at the same pace. It can reveal that short rest intervals are turning a strength session into a conditioning session. It can show that poor sleep, heat, stress, alcohol, dehydration, or illness may be changing your normal response. It can help coaches start better conversations with athletes. It can also help beginners notice that strength training affects the cardiovascular system, even when the goal is muscle or power rather than endurance.

 

The worst role is pretending the watch knows the whole workout. It does not know how heavy the bar felt in your hands. It does not know whether your last rep slowed down. It does not know whether your technique stayed stable. It does not know whether your grip failed before your back, whether your knee felt irritated, or whether you left two reps in reserve. Smartwatch heart rate lifting data is one layer. Training logs, load, repetitions, rest time, perceived exertion, technique quality, pain signals, and long-term progress add the missing layers.

 

For writers and readers interested in wearable fitness data quality, the cleanest summary is this: wrist heart rate is useful when interpreted as an estimate. It becomes risky when treated as a verdict. Smartwatch heart rate accuracy during strength training is affected by sensor physics, algorithm design, wrist movement, contact pressure, exercise type, and the short-stop rhythm of lifting. The research supports trend use more than moment-by-moment certainty. It also supports skepticism toward calorie numbers, especially in resistance exercise.

 

The action plan is direct. Wear the watch correctly. Keep the sensor clean. Use the strength-training mode. Judge trends, not isolated spikes. Consider a chest strap for workouts where heart rate precision matters. Ignore calorie estimates when making nutrition decisions. Record loads, reps, sets, rest periods, and perceived effort. Watch for symptoms that require medical care. Do not let an app score define whether a workout mattered. A watch can count signals; it cannot understand the full cost of a hard set.

 

Disclaimer: This article is for general education only. It does not provide medical advice, diagnosis, treatment, or individualized exercise prescription. People with cardiovascular disease, unexplained chest symptoms, fainting, severe shortness of breath, abnormal palpitations, or medical restrictions should consult a qualified health professional before using wearable data to guide exercise. Consumer smartwatches and fitness trackers are not substitutes for clinical evaluation or emergency care.

 

If you use a smartwatch during strength training, use it as a tool with boundaries. Let it show patterns. Let it support consistency. Let it remind you that the body responds to training in measurable ways. But do not hand it the keys to your judgment. In the weight room, the smartest lifter is not the person who believes every number; it is the person who knows which numbers deserve trust.

 

References

 

Kim KB, Baek HJ. Photoplethysmography in wearable devices: a comprehensive review of technological advances, current challenges, and future directions. Electronics. 2023;12(13):2923. doi:10.3390/electronics12132923

 

Bent B, Goldstein BA, Kibbe WA, Dunn JP. Investigating sources of inaccuracy in wearable optical heart rate sensors. NPJ Digit Med. 2020;3:18. doi:10.1038/s41746-020-0226-6

 

Lee TH, Jun DU, Bae JY, Roh HT, Cho SY. Comparative validity of smartwatch-derived heart rate and energy expenditure during endurance and resistance exercise. Sensors (Basel). 2026;26(8):2526. doi:10.3390/s26082526

 

Fuller D, Colwell E, Low J, et al. Reliability and validity of commercially available wearable devices for measuring steps, energy expenditure, and heart rate: systematic review. JMIR Mhealth Uhealth. 2020;8(9):e18694. doi:10.2196/18694

 

Zhang Y, Weaver RG, Armstrong B, Burkart S, Zhang S, Beets MW. Validity of wrist-worn photoplethysmography devices to measure heart rate: a systematic review and meta-analysis. J Sports Sci. 2020;38(17):2021-2034. doi:10.1080/02640414.2020.1767348

 

Van Oost CN, Masci F, Malisse A, et al. Accuracy of heart rate measurement under transient states: a validation study of wearables for real-life monitoring. Sensors (Basel). 2025;25(20):6319. doi:10.3390/s25206319

 

American Heart Association. Warning signs of a heart attack. Updated December 12, 2024. https://www.heart.org/en/health-topics/heart-attack/warning-signs-of-a-heart-attack