Garmin Pulse Ox Accuracy: A Precise-Looking Number That Should Barely Change Your Training
Share
Here is the honest answer to the question that brought you here. In good conditions — snug strap, still arm, warm hands — Garmin's Pulse Ox gives a reasonable estimate of your blood oxygen saturation. It is less reliable than a fingertip oximeter. And on an ordinary day at sea level, it should change nothing at all about your training.
That last part is the bit nobody writes. Pulse Ox is the number on your watch with the widest gap between how precise it looks and how much it should influence a decision. A crisp two-digit figure feels authoritative; underneath, it is the hardest measurement your watch attempts.
None of this is medical advice. A watch cannot diagnose anything, Garmin does not claim it can, and nothing here is a substitute for a clinician. If you feel unwell or a reading worries you, speak to a doctor rather than interpreting a graph.
What your watch is actually doing
Garmin calls it Pulse Ox. The quantity it estimates is SpO2 — the proportion of the haemoglobin in your arterial blood that is carrying oxygen.
The physics is well established. Oxygenated and deoxygenated haemoglobin absorb red and infrared light differently, so shining both wavelengths through pulsing blood and measuring what comes back tells you roughly how saturated the blood is. That is why a Pulse Ox capable watch has extra LEDs behind the glass alongside the green ones it uses for heart rate.
A hospital oximeter works on the same principle. The difference is where it is pointed.
Why the wrist is so much harder than a fingertip
A fingertip clip works in transmission: light goes in one side and out the other, through thin, well-perfused tissue with a strong pulse. A watch works in reflectance: light has to scatter back out of the same surface, through skin, fat, tendon and bone, from a site with a much weaker pulse. That one difference drags a long tail of problems:
- Fit and position. A loose strap lets ambient light in and lets the sensor move, and worn too far down there is bone and tendon where you want blood vessels. This is the biggest controllable factor.
- Movement. The signal is small, and arm movement swamps it. That is why your watch asks you to hold still.
- Cold and poor circulation. Cold hands mean constricted blood vessels at the wrist, so there is less pulsing blood to read. Readings taken with cold hands are the least trustworthy you will take.
- Ink and skin. Tattoos over the sensor interfere with any optical measurement. Skin pigmentation can also affect pulse oximetry accuracy — a recognised limitation of the technique in general, not something specific to Garmin, and one more reason a wrist number is not a clinical number.
A direction rather than a number, because honest error bands depend on the device, the person and the conditions: expect wrist readings to be noisier than a fingertip clip, and expect the failure to be a spuriously low reading far more often than a high one. Poor signal and poor circulation both push the estimate down. Almost every "my Pulse Ox is alarmingly low" post is a measurement problem.
The precision-to-decision gap
Trusting a reading and acting on one are different questions, and for Pulse Ox the answers are very different.
| What you are looking at | How much to trust it | What it should change |
|---|---|---|
| One spot check at sea level, still, snug, warm hands | Reasonable, for what it is | Nothing. A normal reading is normal, and you already assumed that |
| One spot check that comes out surprisingly low | Suspect the measurement first | Re-take it properly. If it repeats and you feel unwell, that is a conversation with a doctor, not a training decision |
| A single night's graph with dips in it | Low. Position, strap slip and an arm under a pillow all produce dips | Nothing on its own |
| Overnight readings at sea level over weeks | The trend beats any single night | Still almost nothing. It is context, not an instruction |
| Spot checks on arrival at altitude | The metric's best use | How hard to train in the first days up high |
| Daily readings across a week at altitude | Genuinely useful as a trend | Whether you are adapting, and how fast to build |
Read the right-hand column again. Three of the six rows say "nothing" or "almost nothing", one says "go and see a doctor", and the only two that produce a training decision are both about altitude. That is not a criticism of the sensor — it is what blood oxygen is like in a healthy person at sea level. Your body defends saturation hard, so it is one of the last things to move, which is why it makes a poor training signal and a reasonable clinical one.
Contrast that with HRV status and sleep score, which move week to week and do feed Garmin's own training recommendations. Pulse Ox mostly sits on its own, and that is the clue to how much weight to give it.
Where it earns its place: altitude
Go up a mountain and the air pressure falls, so the partial pressure of oxygen falls, so less oxygen crosses into your blood for the same breath. Your saturation drops. That is not a malfunction; it is the reason altitude training exists.
Over the following days your body responds — breathing changes, plasma volume shifts, and over a longer stay red cell production increases. As you adapt, saturation tends to recover toward your own normal. Tracking that recovery is the one thing Pulse Ox does that nothing else on your watch does.
Practically, on a trip up high:
- Take a reading at the same time each morning, in the same conditions, before coffee. Direction over days matters more than any single morning.
- If it is still well below your sea-level normal and you feel rough, keep training easy rather than pushing.
- Some Garmin watches add an acclimatisation readout that combines Pulse Ox with recorded altitude. Treat it the same way: a trend, not a verdict.
Altitude illness is a medical matter, not a data matter. Headache, nausea, breathlessness at rest or confusion at altitude need a person, not a graph.
The battery cost nobody writes about
Pulse Ox is expensive to run. It needs extra LEDs at longer wavelengths, driven harder than the ones used for heart rate, and each reading means seconds of continuous illumination rather than a brief sample. On most Forerunner, Fenix, Venu and Vivoactive watches you can set it to read only when you ask, only while you are asleep, or all day — and those three choices have very different consequences.
| Setting | What you get | What it costs |
|---|---|---|
| Spot checks only | A reading whenever you deliberately take one, which is when you want one | Essentially nothing |
| During sleep | An overnight trend and an overnight low | A real but modest dent in battery life |
| All day | Readings through the day that you will almost certainly never act on | The largest discretionary battery drain on most Garmin watches |
Several Garmin models quote battery life separately for all-day Pulse Ox on their own specification sheets, which tells you how significant they consider it.
Here is the second-order cost, which is the bit that actually matters: a watch that runs flat mid-week is a watch that is not on your wrist overnight. What you lose is your sleep and your overnight heart rate variability — the two measurements that genuinely do drive your recovery and readiness numbers. You would be trading a signal that changes your training for one that does not.
So, plainly: at sea level, leave it off, or sleep-only if you like the data. Switch all-day on for an altitude trip, and off when you get home.
How to take a reading worth looking at
A badly taken reading is worse than none, because you will worry about it.
- Strap snug — firm enough that the watch does not slide, loose enough to be comfortable.
- Worn above the wrist bone, roughly a finger's width up your forearm, not sitting on the joint.
- Arm supported and still, at about the height of your heart, for the whole measurement.
- Hands warm. If they are cold, warm them first and try again.
- Repeat before you believe anything surprising. Two readings that agree beat one that does not.
Overnight readings, and the line this article will not cross
A lot of people arrive here because their overnight graph has dips and a forum thread told them what that means.
Here is where this post stops. A wrist watch cannot diagnose a sleep disorder, and this article is not going to try. Overnight dips have a long list of innocent explanations — sleeping on the arm, the strap loosening, a cold room restricting circulation at the wrist, the sensor losing contact as you turn over. On most watches the measurement is sampled periodically through the night rather than run continuously, so the graph is a series of snapshots, not a recording. Your overnight respiration rate is worth reading beside it rather than instead of it.
What is also true: if you snore heavily, wake unrefreshed, feel sleepy during the day, or someone has noticed you stop breathing in your sleep, that deserves a conversation with a doctor — regardless of what your watch shows. A reassuring graph is not a reason to skip that conversation, and an alarming one is not a diagnosis. Again: not medical advice.
What Pulse Ox does not feed
Garmin's metrics are heavily wired into each other. Sleep, stress, heart rate variability and resting heart rate get recombined repeatedly, so one bad week can appear as four separate warnings.
Pulse Ox is the outlier. As far as I can establish, it is not an input to recovery time, training status or training readiness — it sits to one side, reported but not consumed. The map of which metric depends on which is worth ten minutes; Pulse Ox has the fewest arrows pointing out of it.
If you want to spend attention on a watch number that really does change your training, spend it on your heart rate zones. A maximum heart rate left at the age-formula default mislabels every session you do, for as long as it stays wrong. That is a better use of an afternoon than a blood oxygen graph.
The wider point
Your watch measures well and advises badly, and Pulse Ox is the clearest example in the range. It produces a precise-looking figure and leaves you alone with what to do about it — which, most days, is nothing.
Closing that gap means knowing which numbers deserve to change a session and which are just scenery. Lactate Training syncs your Garmin activities and health data and uses them to build and adapt a plan, with zones anchored to a measured lactate threshold rather than an age formula, so the data that should change your week does and the data that should not is left alone. However you do that job, the numbers are inputs, not instructions.
Frequently asked questions
Is Garmin Pulse Ox accurate?
It is a reasonable estimate in good conditions — snug strap, still arm, warm hands — and noticeably less reliable than a fingertip oximeter. The wrist is a hard place to measure blood oxygen, because the sensor has to read light scattered back out of thick, less well-perfused tissue. Garmin does not present it as a medical device.
Why is my Garmin Pulse Ox reading low?
Almost always because of how it was taken. A loose or badly positioned strap, a moving arm, cold hands or a tattoo under the sensor all push the figure down. Re-take it with the watch snug and above the wrist bone, your arm still and your hands warm. If a properly taken reading is repeatedly low and you feel unwell, speak to a doctor.
Should I turn Pulse Ox on all day?
For most people at normal altitude, no. All-day tracking is the biggest discretionary battery drain on most Garmin watches and produces readings you will almost never act on. Sleep-only is a fair middle ground, and all-day earns its place on a trip to altitude. A flat battery costs you the sleep and heart rate variability data that does drive your recovery numbers.
Does Pulse Ox drain the Garmin battery?
Yes, significantly, and more than most people expect. It uses extra LEDs driven harder than the heart rate sensor, and every reading means seconds of continuous illumination. Several Garmin models quote a separate, shorter battery figure for all-day Pulse Ox on their own specification sheets.
Can my Garmin detect sleep apnoea?
No. A wrist watch cannot diagnose a sleep disorder or any other condition, and dips in an overnight graph have many innocent causes, including sleeping on the arm and the strap loosening. If you snore heavily, wake unrefreshed or someone has noticed you stop breathing in your sleep, raise it with a doctor regardless of what your watch shows. This is not medical advice.
Does Pulse Ox affect my training readiness or recovery time?
As far as can be established, no. Unlike sleep, stress and heart rate variability, which are recombined into several of Garmin's recovery metrics, blood oxygen appears to be reported on its own rather than consumed by the training numbers.
Spend your attention on the numbers that change a session
Blood oxygen is scenery; your zones, your load and your sleep are the data that should actually move a training week. Lactate Training reads your Garmin activities and health data and adapts the plan around the signals that matter, with zones built on a measured lactate threshold instead of an estimate.