Heart rate training has anchored endurance programming for 40 years, and the basic principle remains sound: relative cardiovascular intensity, measured as a percentage of maximum heart rate or heart rate reserve, predicts metabolic stress with reasonable precision at steady-state efforts. The problem is that the accuracy of the measurement itself varies enormously depending on the hardware you are using, the conditions, and the type of exercise. Training with inaccurate heart rate data in structured zones is worse than training without zones — you are making confident decisions based on wrong numbers.
The commercial wearable market has shifted dramatically toward optical wrist sensors because they are more comfortable, require no chest strap, and have improved substantially since 2018. But "improved" is relative. The gap between a quality ECG chest strap and a quality optical wrist sensor remains significant at exercise intensities above Zone 2 aerobic, and there are specific conditions — cold weather, high-intensity intervals, dark skin pigmentation — where optical sensors fail systematically and predictably. Understanding where each technology is reliable allows you to use them appropriately rather than dismissing one or the other entirely.
How ECG Chest Straps Work
Electrocardiography-based chest straps detect the electrical potential generated by cardiac depolarisation — the same signal a clinical ECG reads. Two electrodes on the strap make skin contact across the chest wall, and the device detects the R-wave peak of each heartbeat, calculating instantaneous heart rate from the time between consecutive R-waves (R-R interval). Because the signal is the actual electrical event of cardiac contraction, there is essentially zero lag between a heartbeat and the recorded value — response time is within a single heartbeat cycle.
Accuracy of quality chest straps (Polar H10, Garmin HRM-Pro, Wahoo TICKR X) in research conditions is exceptional: mean absolute error (MAE) of 1–2 bpm across most exercise conditions, with R² values of 0.99 versus laboratory ECG reference in multiple validation studies. The primary sources of error are poor electrode contact (dry skin, poor fit, synthetic fabrics that reduce conductivity) and electrical interference from power transmission lines — both are easily controlled. Chest straps are the reference standard for running heart rate research for this reason.
How Optical Wrist Sensors Work — and Where They Fail
Photoplethysmography (PPG) optical sensors use green LED light (480–560 nm wavelength) shone into the skin at the wrist. Oxygenated haemoglobin in blood vessels absorbs green light more strongly than deoxygenated haemoglobin, so as blood pulses through capillaries with each heartbeat, the light reflected back to the sensor oscillates. Algorithms detect this pulsatile signal and extract heart rate. The core challenge is that PPG is a volumetric blood flow measurement, not an electrical one — it depends on good optical contact, stable wrist positioning, and sufficient perfusion at the measurement site.
Research has quantified the failure modes precisely. A 2019 study in the Journal of Medical Internet Research testing the Apple Watch Series 4, Garmin Fenix 5, and Fitbit Charge 3 against Polar H10 found mean absolute errors of 3.2–5.8 bpm during steady-state running — acceptable for Zone 2 work. During high-intensity interval running (6 × 1 km at 5 km race pace), MAE rose to 11.4–17.7 bpm, with peak instantaneous errors exceeding 25 bpm. At 170+ bpm intensities, wrist sensors frequently lock onto harmonic frequencies of wrist movement (180 steps per minute ≈ 180 bpm) rather than the cardiac signal — a well-documented artefact called motion artefact contamination.
- Cold conditions (<10°C): peripheral vasoconstriction reduces wrist blood flow, weakening the PPG signal; errors of 8–15 bpm are common in sub-zero conditions
- Dark or heavily tattooed skin: melanin absorbs green LED wavelengths, reducing signal-to-noise ratio; MAE increases 3–4× versus light skin in multiple studies; some devices now offer red/infrared LED alternatives with improved accuracy for darker skin tones
- Wrist hair density: thick forearm hair between the sensor and skin degrades optical contact; wear the sensor 2 cm above the wrist bone and snug (not tight) for best results
- Running form and arm swing: high-amplitude arm movement at the wrist increases motion artefact; structured interval workouts with aggressive arm drive show higher error rates than steady tempo running
Accuracy Data and What It Means for Training Zones
The practical consequence of measurement error depends on your training zone boundaries. Typical Zone 2/3 boundary is placed at 75–80% of maximum heart rate — for a runner with HRmax of 180 bpm, that is 135–144 bpm. A ±5 bpm MAE during steady running means a wrist sensor could place you in Zone 3 when you are in Zone 2, or vice versa, on any given day. Over a training block of 3–4 runs per week, this is meaningful accumulated training load misassignment.
At threshold intensity (Zone 4, 88–92% HRmax), the stakes are higher. A 10–15 bpm overread during an interval session can trick an athlete into believing their threshold pace is slower than it actually is — leading to undertrained intervals and an inaccurate lactate threshold estimate. Conversely, a 10 bpm underread during a threshold run encourages the athlete to push harder to "hit" the zone, creating unplanned overreaching. For athletes using heart rate for threshold and VO2max sessions, chest strap accuracy is not optional — it is the foundation of the entire training structure.
When to Use Each Technology: Recommendations by Use Case
The practical decision tree is straightforward. For easy and recovery runs where Zone 1–2 accuracy of ±5 bpm is acceptable and the primary goal is keeping intensity low, optical wrist sensors are adequate and the convenience advantage is real. For threshold, VO2max, and race-intensity sessions where accurate zone placement matters, a chest strap is the correct tool. Racing is a grey area — comfort and practicality favour wrist sensors for most athletes in competition, but understanding that race-day wrist data may be noisy above 85% HRmax is important for post-race analysis.
For athletes doing HRV-based training (heart rate variability), there is no ambiguity: only ECG chest straps provide R-R interval data with the 1–2 ms precision required for meaningful HRV analysis. Optical sensors reconstruct R-R intervals from a smoothed waveform and typically introduce 3–8 ms of additional noise — sufficient to distort HRV metrics by 10–20%. The Polar H10 is specifically validated for HRV research purposes, and its accompanying app (Elite HRV, HRV4Training) integrates directly with it for morning readiness protocols.
- Easy / Z1–Z2 steady runs: optical wrist sensor is adequate; review data post-run for any clear artefacts
- Threshold and interval sessions: ECG chest strap strongly recommended; zone accuracy matters
- Cold weather running (<5°C): chest strap; optical sensors unreliable due to vasoconstriction
- HRV morning readings: ECG chest strap only (Polar H10 is the research standard)
- Swimming: some chest straps store data; optical sensors are entirely unusable in water
- Cycling (stationary or road): optical sensors perform better on the wrist due to reduced motion artefact versus running; still less accurate than chest strap above threshold
Improving Optical Sensor Accuracy: Practical Steps
If a chest strap is unavailable, several steps meaningfully improve optical sensor performance. Wear the device snugly (it should not spin on the wrist) and position it 2 cm proximal to the wrist bone — further from the joint reduces motion artefact. Allow 2–3 minutes of easy jogging before any intensity work to let the sensor stabilise its baseline signal. In cold conditions, wear a thin liner glove that covers the wrist sensor, trapping warmth and preserving peripheral perfusion. Post-session, review the HR trace for implausible spikes or flat-line dropouts and cross-reference with RPE to identify sessions where the data is not trustworthy.
Some newer devices (Garmin Forerunner 965, Apple Watch Ultra 2) use dual-wavelength PPG (green + red LED) and enhanced motion compensation algorithms that reduce running MAE to 4–6 bpm — better than earlier generations, but still below ECG accuracy at high intensity. The manufacturers' claims of "medical grade" accuracy refer to resting conditions and should not be extrapolated to interval running scenarios.
Accurate heart rate data is only as useful as the zone system it feeds into. Use the NorthLine Training Load Calculator to translate your heart-rate-based training data into acute and chronic load metrics, ensuring your zone-based work accumulates into appropriate progressive overload without overreaching.