Sports watches and heart rate monitors: how accurate they really are

You are halfway through an interval, you glance at your watch and it reads 171. The person next to you is wearing a chest strap and their screen says 154 at the same moment. Neither device is faulty — they measure heart rate in completely different ways, and one of them is guessing far more than the other.

This is a buying guide, not a review of any particular model. The point is to work out which readings deserve your trust and which features are worth paying for. One rule covers most of it: the further a number sits from the sensor that produced it, the bigger the error.

What the device measures and what it calculates

A sports watch genuinely measures very few things: heart beats (optically, or through electrodes), position from satellites, wrist acceleration, air pressure for altitude, and sometimes skin temperature. That is the whole list.

Everything else is calculated from those inputs using a formula the manufacturer does not publish: calories, estimated VO2 max, training zones, stress level, sleep stages, lactate threshold, readiness scores. Those numbers are not invented, but they are estimates built on a stack of assumptions about you.

How accurate each reading actually is

Heart rate at rest and in steady effort

This is where a wrist sensor does well. Sitting still, jogging at an even pace or spinning on a stationary bike, the gap between wrist and chest strap usually stays within a couple of beats. A chest strap reads the electrical signal of the heart and is treated as the practical reference point.

Heart rate during intervals and lifting

This is where it falls apart. An optical sensor reads changes in blood flow under the skin, so it needs roughly 10 to 30 seconds to catch a sharp rise. On short intervals that means the watch shows your peak heart rate only after you have already stopped.

The other classic failure is cadence lock: the sensor latches onto the rhythm of your steps instead of your pulse and calmly displays a rock-steady 160 while you are actually at 130. Lifting is worse still, because gripping hard tenses the forearm, squeezes the capillaries and shifts the watch. Holding an Olympic barbell tightly or swinging a kettlebell can produce errors of 20 to 40 beats through parts of a set. For that kind of training a chest strap or an armband sensor solves the problem outright.

GPS and distance

Outdoors with a clear view of the sky, modern receivers usually keep distance error somewhere between one and three per cent. Between tall buildings, under dense tree cover or in a narrow valley that error can multiply several times over, because the signal arrives bounced rather than direct.

Instant pace

Current pace is the least stable number on the whole watch — it can swing by 20 to 30 seconds per kilometre in either direction. Lap pace or per-kilometre pace is far steadier, and that is the one you should be running to.

Step count

Daily steps are typically within about 5 to 10 per cent, but only when you walk normally. Pushing a trolley, carrying a bag in one hand or driving on a rough road produces both phantom steps and missed ones.

Calories

The weakest figure any of these devices produce. Energy expenditure is estimated from heart rate, weight, age and sex, while real expenditure also depends on fitness, movement efficiency and the type of effort. Errors of 20 to 30 per cent are common and strength sessions can be further out than that. If you are tracking food intake, do not feed this number into the sum.

Sleep, heart rate variability and readiness

Devices are reasonably good at spotting when you fell asleep and when you got up. Splitting the night into stages is much shakier and the evidence does not agree. Heart rate variability is worth following as a trend across weeks, measured under the same conditions each time — a single morning tells you nothing useful.

A score like "readiness 62" is not a measurement, it is a summary of estimates. Treat it as rough confirmation of what you already feel, not as an instruction.

What the device genuinely has to have

• The ability to pair with an external heart rate sensor. It is the one feature that guarantees accuracy when you need it.
• Battery life that covers your longest session plus several ordinary days.
• A screen you can read in sunlight while moving, with digits large enough that you do not have to stop.
• Data export in a standard format and syncing with the app you already use.
• Water resistance of at least 5 ATM if you swim or run in the rain.
• A strap you can tighten enough that the watch does not rock about. Without that, the optical sensor has no chance.

The specifications that matter, and what they mean

Sensor type

An optical sensor sits on the wrist and reads light returning from the skin. Chest electrodes read the heart's electrical signal, which is why they respond faster and hold steadier. An armband sensor is the middle ground: worn like a strap but on the arm, and clearly better than a wrist watch under load.

Connectivity

Bluetooth is supported almost everywhere. ANT+ matters if you want one sensor broadcasting to your watch and a bike computer at the same time. Check whether the watch can also broadcast your heart rate to gym machines.

Satellite receiver

A single-frequency receiver is plenty for roads and open ground. Dual-band helps noticeably in cities, forests and mountains, but it costs both battery life and money.

Battery

Read the two figures separately: hours with GPS running, and days in everyday use. Across the market that runs from roughly 6 hours to well over 40 in GPS mode, and from 3 to 20 days otherwise. If your sessions last an hour, do not pay for ultramarathon endurance.

Materials and weight

A plastic case with mineral glass is lighter and cheaper; metal and harder glass survive knocks and scratches better but add mass. Sports watches run roughly 30 to 90 grams, and you feel the difference if you sleep in one.

Case size

A large case on a slim wrist sits unstable, and an unstable watch means poor heart rate data. If your wrist measures under about 16 centimetres, look for the smaller version of the same model.

Software and your data

Before buying, check that you can export your training history and take it elsewhere. You will replace the watch in a few years; the history should outlast it.

What you actually need to spend

Think in tiers rather than exact figures. A chest strap is the cheapest accuracy upgrade available and costs a fraction of any watch.

An entry-level GPS watch with heart rate gives you distance, pace, pulse and zones, which is around 90 per cent of what an ordinary runner needs. The mid tier adds better battery life, a sharper receiver, more sports and more derived metrics. The top tier is where you pay for materials, maps, dual-band GPS and multi-day endurance.

A rough rule: if you train three to five times a week on familiar routes, jumping from mid to top tier will not change your training. Adding a chest strap will.

Common buying mistakes

• Paying for metrics you will never open. Wrist-based lactate threshold sounds impressive until you notice you never look at it.
• Building your food intake around the calorie figure.
• Wearing the watch too loose, or too close to the hand. Move it two fingers above the wrist bone and tighten it.
• Comparing two devices second by second and hunting for the "correct" one. Compare the average and the shape of the curve across the whole session.
• Ignoring the app. You will spend more time in it than with any button on the watch.
• Buying the largest model for looks, then getting poor readings because it never sits flush against the skin.

Who it pays off for, and who it does not

It pays off if you run, cycle, swim or do intervals and want to keep effort inside sensible limits rather than guessing every session. It pays off if you care about trends across months: resting heart rate, weekly volume, how long recovery takes. If you are building a smarter plan in your fifties, that trend line matters more than any single workout, and there is more on that in the piece on what smart training at 50 looks like.

It does not pay off if you mostly lift and track sets, reps and kilos — a notebook or a plain logging app does the same job, and wrist heart rate is unreliable there anyway. It also does not pay off if the number on your arm turns into pressure and you sleep worse because of a sleep score.

When to see a doctor

A watch is not a diagnostic tool, and a wrist ECG feature is not a substitute for being examined. But if the device keeps flagging an irregular rhythm, if your resting heart rate climbs without explanation and stays up for weeks, or if your pulse refuses to settle long after an easy session, book an appointment.

Chest pain, fainting or heavy breathlessness during efforts you used to handle comfortably are not something to argue about with a wrist reading — get them checked.

The short version

Heart rate at rest and in steady effort: reliable. Heart rate in intervals and under load: unreliable without a strap. Distance: good in the open, poor among buildings. Calories: rough guidance only. Sleep stages and readiness scores: interesting, not measurements.

Buy a device that does the basics well, add a chest strap if you train hard, and let the rest be a bonus.