Swimming and Heart Rate: Why Your Apple Watch Loses the Plot in the Pool

You know exactly how a hard 100 feels. Your breathing is loud, your arms are burning, you push off the wall and your heart is clearly working. Then you glance at your Apple Watch and it says 112 bpm. A minute later it says 96. At no point does it agree with what your body is telling you.

This is one of the most common complaints from swimmers who train with a watch, and it is almost never about fitness. It is about physics. Wrist-worn optical heart rate sensors have a genuinely hard time in water, and there is a solid body of research showing exactly how far off they can be.

TL;DR: Wrist optical heart rate is the least reliable part of your swim data. PPG sensors need light to bounce off blood flow in your skin, and in the pool three things break that: a film of water between the sensor and your skin, cold-water vasoconstriction that reduces blood flow in your hands and wrists (a factor Apple itself documents), and constant arm movement plus water impact at the entry phase. Research on front crawl swimming found wrist optical readings with a mean bias of about −26.5 bpm and limits of agreement from −73 to +20 bpm against a chest strap, while an optical sensor on the temple or goggles tracked the chest strap almost perfectly. Apple Watch's own answer is to lock the screen, track your laps, sets and rest automatically, and keep recording. It cannot magically fix optical sensing underwater. So treat swim heart rate as a floor, not a measurement: use pace per 100m and RPE as your primary intensity markers, check that your resting heart rate and HRV respond the next morning, and if you want swim zone training to actually work, add a chest strap or a non-wrist optical sensor.

YouTube: Related video

Why water is such a hostile environment for an optical sensor

Your watch does not measure heart rate directly. It shines light into your skin and measures how the light scatters as blood pulses through the vessels underneath. That requires two conditions: consistent contact between the sensor and your skin, and enough blood flow near the surface to create a detectable signal.

Water attacks both.

1. The fluid film between sensor and skin. Water gets under the watch during the entry phase, at the turns and during the catch. Even a thin layer changes the optical path between the LEDs and the photodiodes. In swimming studies, researchers consistently point to the fluid film as one of the main reasons wrist readings degrade, and the strap tension changes as the arm moves — sometimes loosening, sometimes pressing harder, both of which distort the signal.

2. Vasoconstriction and cold. Apple's own technical paper on heart rate measurement states that skin perfusion varies significantly between people and can be affected by the environment, noting that when users exercise in the cold or swim in a pool, blood vessels may undergo vasoconstriction, especially in the extremities. Less blood flow at the wrist means a weaker pulse signal to detect. This is a documented limitation from the manufacturer, not a conspiracy theory from disappointed swimmers.

3. Arm movement and impact. Front crawl is about as motion-heavy as it gets. The arm accelerates, rotates, enters the water and pulls, over and over, while the sensor is trying to detect a signal of a few percent amplitude. Motion artifacts and the physical impact of the hand hitting the water are both recognised problems in the literature.

Other studies have shown that accuracy drops as intensity rises — which is the worst possible direction of error for a swimmer trying to train intervals, because the harder you go, the more likely the watch is to under-read the effort.

What the research actually shows

This is not a vague gripe. There are several controlled studies using a chest strap as the reference (chest straps read the heart's electrical signal, and are treated as the practical gold standard) and comparing them with wrist watches during front crawl.

Read those numbers again, because they matter for how you interpret your watch. A mean bias of −26.5 bpm is not a rounding error. If your true heart rate is 150 in a hard 100, a reading of 123 is entirely consistent with the published data. And the limits of agreement tell you the individual readings can be off by far more than the average — one study's wrist data spanned roughly 90 bpm between the upper and lower agreement limits.

Two useful takeaways from the same body of work:

• Placement matters more than brand. Every study that put an optical sensor somewhere other than the wrist — temple, goggles, occasionally upper arm — got dramatically better results in water. The problem is the wrist being in the water and moving, not the technology itself. On land, these same watches perform very well.
• Temple placement wins. Sensors held against the temple under a swim cap showed near-perfect agreement with chest straps. Head movement in swimming is far smaller than arm movement, pulse pressure is higher there, and the cap keeps a stable, water-free interface.

So what is your watch actually showing you?

Three distinct failure patterns, and they mean different things.

Low readings that drift down. The classic. The sensor is picking up something, but not your pulse. What you often see is a reading that was correct during the warm-up, then drifts as the sensor loses and slowly reacquires. This is the pattern the studies describe, and it is the one that tempts you into thinking swim intensity is impossible to measure.

Gaps. Nothing recorded for stretches, or a flat line. This happens when the sensor cannot find a signal at all — often right after a turn, during the fastest intervals, or when the strap is loose. Gaps are honest: the watch is telling you it does not know.

Occasionally, plausible-looking data. This is the dangerous one. A swim file that looks smooth and sensible may still be several beats off. There is no way to tell from the graph alone. If your watch happens to return a clean curve, you still should not compare it directly with your running numbers.

It is worth saying clearly: if you see HR values in the 90s during a hard set, this is not data about your heart. It is data about your sensor. Both the general accuracy picture for wrist optical sensors and the swimming-specific studies point the same way.

The zones problem, which is bigger than the sensor problem

Even with perfect heart rate data, swim zones are not land zones.

Your watch computes heart rate zones from an estimate of your maximum heart rate, which is generally derived from land activity, and Apple uses the same zone framework across sports. But maximal heart rate during swimming is typically lower than during running for most people. Coaches and sports scientists commonly cite a swimming HRmax roughly 10 to 15 bpm below a runner's, because of the horizontal body position, the cooling effect of the water and the smaller active muscle mass relative to running.

Put the two problems together and you get the familiar swim training result: a genuinely hard set that would be Zone 4 on the bike lands in Zone 1 or 2 on your wrist, both because the sensor under-reads and because the zone boundaries were built for running.

If you want to work with zones anyway, the practical fix is to build your own swim-specific zone map over a few weeks. Swim a controlled set — say 4 × 200m at a pace you can just about hold conversationally, then 4 × 100m at a pace your breathing makes talking hard — and record the RPE, the pace per 100m and whatever heart rate the watch reports. The absolute HR numbers will be unreliable, but the relative ordering usually holds, and pace per 100m becomes your real intensity anchor.

How to get better swim heart rate data

You cannot fix the physics from your wrist, but you can improve the odds.

Before you get in:

• Wear the watch higher. Apple recommends placing it on the top of the wrist, snug against the skin, above the wrist bone rather than over it. On a wrist that flexes constantly during a pull, position drift of a centimetre can be the difference between a signal and nothing.
• Tighten by one notch compared with your daily fit. The band should not rotate on the wrist, especially during the catch.
• Clean the back of the sensor. Sweat, sunscreen and pool residue all sit between the LEDs and your skin.
• Start the Pool Swim workout before you enter — the workout automatically turns on Water Lock so the screen stops responding to water taps, and it starts tracking sets and rest. This matters because the watch uses lap and set structure along with motion data to make sense of a swim (Apple explains the setup in its guide to swimming with Apple Watch).

In the water:

• Stay in during rest intervals. Sensor reacquisition after getting out and back in usually costs you the first part of the next interval.
• Favour longer intervals for measurement. On a 3 × 100m set a large fraction of the session is transition, and transitions are where the data is worst. If you want a comparable number across weeks, use sets of 200m or more and look at the steady middle.
• Use RPE as your primary scale and let heart rate be a secondary check. On a 1 to 10 scale, a solid aerobic swim is 6, threshold work with hard breathing that you can sustain is 7 to 8, and anything above 9 should be short.
• Watch the recovery, not only the effort. Your heart rate recovers in a clear pattern after hard intervals: it drops quickly in the first minute or two, then more slowly. Even a partially reliable curve still shows that shape, and if it disappears for weeks, that is a genuine signal about fatigue.
• Add a second sensor if you care about swim intensity. A Bluetooth chest strap paired to your watch is the cheapest route to real swim heart rate, and the studies are unambiguous that chest straps remain the reference during swimming. If a strap is uncomfortable or slides, an optical sensor worn at the temple under the cap, or on the upper arm, both outperform the wrist — with the temple clearly the better of the two in water.

After the swim:

• Look at the next morning, not the swim graph. Resting heart rate and HRV the day after a hard swim tell you whether the session was genuinely taxing in a way the pool-side numbers usually cannot. If your resting heart rate is up several beats and your HRV sits below your baseline, the session did what you thought it did. That is the same logic behind using a single daily recovery picture instead of chasing one workout metric — a daily health score, recovery score and sleep trend built from overnight data, so you are reading patterns rather than trusting one noisy curve.

Three quick tests to find out how bad your data is

1. The dry-land control. Record a 20-minute run or bike with your watch, then a pool session with the same strap tightness. If land data is clean and swim data is nonsense, it is water and motion, not your watch.
2. The drift test. Swim a 400m at a steady, controlled pace. If heart rate climbs and then collapses to implausible values, you are seeing sensor loss, not physiology.
3. The pull-buoy comparison. Swim the same distance with a pull buoy between your legs. That removes kick and reduces body roll, giving the wrist a more stable position. If your readings suddenly become sensible, movement was the main culprit.

When swim heart rate data is worth trusting

Rarely in absolute terms, and often in relative terms.

• Trust the trend across weeks. If the same set at the same pace shows a consistently lower average heart rate over two months, that is meaningful even if the absolute number is wrong, because the error is roughly consistent.
• Do not trust single readings. A 100 bpm reading after a maximal 50 is not your heart rate.
• Do not use swim heart rate to compare with cycling or running zones. Different sport, different maximum, different sensor quality.
• Do trust the overnight data. Sleeping heart rate, HRV and sleep quality after swim training reflect real physiological load, and they are measured under ideal conditions: still, warm and dry.

Swimming is one of the best things you can do for your aerobic base and your joints, and no wristwatch sensor is going to change that. What it will do is under-report your effort while making you feel like you are not training hard enough. Once you know that the under-reading is systematic and expected, the number stops being confusing and becomes one input among several.

Bottom line

Your Apple Watch is not broken and you are not slacking. Wrist optical heart rate is unreliable in water for documented physical reasons: a fluid film between the sensor and your skin, cold-induced vasoconstriction in the extremities that Apple itself acknowledges, and heavy arm motion during the pull. Published comparisons during front crawl put wrist readings off by an average of tens of beats per minute, with individual readings far worse, while sensors placed at the temple track chest straps almost perfectly.

Three things follow from that. First, treat swim heart rate as a floor rather than a measurement. Second, make pace per 100m and RPE your primary intensity tools, and drop swim zones by roughly 10 to 15 bpm relative to your running ones. Third, if you want real numbers, add a chest strap or a non-wrist optical sensor — because the sensor placement, not the brand, is the thing you can actually change.

Century AI helps you understand your body with a daily health score, recovery score, and sleep insights — using the watch you already wear.