Apple Watch swim tracking: every metric explained, and how accurate it really is

Apple Watch is genuinely excellent at some swim metrics and genuinely unreliable at one of them. Knowing which is which is the difference between data you can train on and numbers you should quietly ignore.

Last updated 11 August 2026 · 9 min read

What Apple Watch measures well

Almost everything that comes from motion rather than physiology. Lap counting, stroke type, stroke count, split times, pool distance, SWOLF and total time are all reliable, and for most swimmers that is the entire useful dataset.

The watch works out what you are doing from its accelerometer and gyroscope, not from a sensor pressed against your skin, and that turns out to be the right tool for the job. A tumble turn or a touch-and-push has a distinctive signature (rapid deceleration, rotation, then acceleration away from the wall) that is easy to detect and hard to fake. Each of the four competitive strokes has a distinctive arm path, so the watch can label freestyle, backstroke, breaststroke and butterfly on a length-by-length basis, and it is right the overwhelming majority of the time for swimmers with reasonably conventional technique.

From those two things, lap detection and stroke recognition, everything else in a pool swim follows arithmetically: distance is lengths multiplied by pool length, pace is distance over time, stroke count comes from counting arm cycles, and SWOLF is the length time plus the stroke count. None of it depends on a sensor reading through water.

Pool mode vs open water mode, and why the difference matters

Pool mode counts laps by detecting turns and multiplies by a pool length you enter, so that number has to be right. Open water mode has no lap length at all, because there are no walls, and estimates distance a different way.

This is the most consequential setting on the watch, and the one most often got wrong. In pool mode the watch is not measuring how far you swam. It is counting how many times you turned around and trusting you about the rest. If you tell it the pool is 25 metres when it is 25 yards, every distance in the session is inflated by about nine percent, your pace per 100 looks better than it was, and nothing on screen warns you. Do that consistently and you will build a training history that is quietly wrong, then set pace zones on top of it.

The error runs the other way too. Some hotel and gym pools are 20m, 33m or an odd imperial length, so a swim there with the watch still set to your home pool produces a plausible-looking session that never happened.

Open water mode drops the concept of a length entirely, and this is deliberate rather than a missing feature. With no walls there are no turns to count, so there is nothing to multiply. Setting a pool length for an open water swim is not just unhelpful, it is actively harmful: the watch will read changes in your motion (sighting, a wave, a change of direction round a buoy) as turns and fabricate phantom lengths. The distance you end up with is not an approximation of your swim, it is an artefact of your stroke.

It also means open water pace and pool pace are not comparable, and should not be pooled into the same averages. Current, chop, sighting and the total absence of push-offs make an open water 100 a different unit of measurement. This is why SwimTune records open water swims with no pool length assigned, and deliberately excludes open water pace from CSS calculations and personal-best records. If you are building toward an open water event, the open water training guide covers how to structure that work.

Every swim metric, and what it is actually for

There are roughly a dozen numbers in a recorded swim, and each answers a specific question. Knowing the question is what stops a metrics screen from being wallpaper.

Worth noting: the recovery metrics in that last row are among the most reliable data your watch produces, precisely because they are collected overnight on dry skin. If you want a physiological signal you can actually lean on, it is far more likely to come from resting heart rate and HRV trends than from a heart rate figure captured mid-stroke. Training load and recovery goes into how to read them together.

Heart rate accuracy in water: the honest answer

Wrist-based optical heart rate is materially less reliable underwater than on land, on every brand of watch. Treat in-water heart rate as indicative of effort, not as a precise measurement, unless you are wearing a chest strap.

Optical sensors work by shining light into the skin and reading how much comes back as blood pulses through the capillaries. Swimming attacks that method from three directions at once:

• Water between the sensor and the skin. Even a thin film changes how light scatters, and a watch that lifts slightly on the recovery lets water in with every stroke.
• Wrist flexion and motion. The wrist bends and rotates through every stroke cycle, and the watch shifts against the skin. That produces motion artefacts exactly where the sensor needs stillness.
• Cold-induced vasoconstriction. Cool water pushes blood away from the periphery to protect core temperature. Less blood near the surface of the wrist means less signal to read, which is why open water readings are usually worse than pool readings.

This is physics, not a design flaw specific to Apple. Garmin watches struggle with the same problem, and independent testing has repeatedly found wrist-based readings under-reading substantially during open water swims, sometimes badly enough that the trace bears little resemblance to a chest strap recorded at the same time. Recent hardware has improved things: the Apple Watch Ultra series has tested notably better than older models, and a well-fitted watch on a warm indoor pool swim can produce a usable trace. But "usable" is not "reference standard".

If you genuinely train by heart rate, wear a chest strap. Practically, that means one with onboard memory that records the session and syncs afterwards, since Bluetooth does not travel well through water and a live connection from chest to wrist mid-length is unreliable. If you do not train by heart rate, and most swimmers do not, you lose almost nothing by treating the in-water number as a rough effort indicator and pacing by the clock instead.

Why swimmers should pace by pace, not heart rate

Because heart rate is the weakest signal available in the pool and pace is the strongest. Even with a perfect sensor, swimming heart rate is suppressed and lagging, while the pace clock is exact and instant.

Swimming heart rate runs lower than running or cycling heart rate at the same relative effort, typically by a meaningful margin, and the reasons are structural. You are horizontal, so the heart is not fighting gravity to return blood from the legs. Water pressure on the chest and limbs assists venous return. The water is cooler than air, so less cardiac output is diverted to cooling the skin. And the working muscles are largely upper body, with a smaller total muscle mass recruited than in running. A "zone 4" number borrowed from your running profile simply does not mean the same thing in the pool.

Heart rate also lags. It takes 30 to 60 seconds to reflect a change in effort, which is most of a 100m repeat. By the time the number tells you that you went out too hard, you have already gone out too hard.

Pace has none of these problems. It is directly measured, it responds instantly, it is comparable across weeks and years, and no sensor artefact can corrupt it. The standard way to anchor it is a threshold test: swim a 400 and a 200 with full recovery between, and derive your critical swim speed, then build five pace zones around it. That gives you a target time per 100 for recovery, aerobic, threshold, VO2max and sprint work, which is a far more actionable instruction than a heart rate band you cannot read while your face is in the water.

Setup that fixes most bad swim data

Nearly every complaint about swim tracking traces back to one of four things: wrong pool length, loose strap, wrong workout type, or a screen touched mid-swim. All four take seconds to prevent.

• Set the pool length before you start, every time. Not once, every time. It is the highest-consequence setting on the watch and it silently corrupts distance and pace when it is wrong. If you swim in yards, set yards.
• Wear the watch snugly, and slightly higher up the wrist. A loose watch is the most common cause of bad data, full stop. Move it a finger's width above the wrist bone and tighten one hole more than you would on land. It should not slide when you shake your arm.
• Let Water Lock engage. It turns off the touchscreen so a stray hand or the water itself cannot pause your workout or skip a set. Starting a pool or open water workout enables it automatically.
• Eject the water afterwards. Turn the Digital Crown to unlock and let the watch clear the speaker. Skipping this does not damage anything, but it leaves the speaker muffled.
• Pick the right workout type. Pool swim for pools, open water for anything without walls. Never open water in a pool, and never pool in a lake.
• Be aware of long push-offs on short pools. If you push off and glide a long way before taking a stroke, especially in a 20m or 25 yard pool, stroke detection has less to work with and stroke counts can come in low. Consistency matters more than perfection here: a small systematic bias still gives you a usable trend.

Do you need a third-party swim app?

No. The built-in Workout app records swims accurately and is free, and for a lot of swimmers that is the whole job. A third-party app is only worth it if it does something the built-in one does not.

It is worth being plain about this. Apple's own app counts your lengths, labels your strokes, times your splits and writes it all to Apple Health. If what you want is a reliable record of the swim you just did, you already have it, and paying for an app that shows the same numbers in a different colour is a waste.

The case for something more starts when you want the data to tell you what to do next: trends across months rather than a single session, structured sets rather than "swim for 40 minutes", pace zones derived from your own threshold test, training load measured against your recovery, and a plan that changes when your life does.

SwimTune is built for that second job. It reads the same HealthKit data your watch already records, so nothing about how you swim changes, then adds the analysis layer: per-length SWOLF and distance per stroke trends, CSS-based pace zones, training load against sleep, HRV and resting heart rate, and adaptive plans that back off when your readiness drops. Recording, importing and full history and analytics are free; the AI-generated adaptive plans and coach chat are the paid tier.

Garmin and other watches

Swims recorded on a Garmin work with SwimTune today, provided Garmin Connect is writing your workouts to Apple Health. They import with the correct venue, distance and splits, and sit alongside Apple Watch swims in the same history.

Everything in this guide about pool length, lap detection and wrist heart rate applies to Garmin, Polar, Suunto and the rest as well. The underlying methods are the same across the category: accelerometer-based turn detection, a pool length you have to set yourself, an optical sensor that fights the same water and cold that Apple's does. If you have been getting reliable lap counts from a Garmin, you will get reliable lap counts from an Apple Watch, and if your Garmin heart rate looks implausible on open water swims, an Apple Watch will not magically fix it.

A native Garmin app for SwimTune is in development and planned for fall 2026. It does not exist yet, so today the Apple Health route is the way Garmin swims get in. Wear OS and Health Connect are further out again, on the roadmap after Garmin.

Common questions