Heart Rate Variability: What Your Watch Is Really Telling You

A heart at 60 beats a minute does not beat once every second. It beats at 0.94 seconds, then 1.08, then 0.97. That wobble is heart rate variability, and a watch now reports it to millions of people every morning with no explanation attached.

What the wobble is#

Two branches of the autonomic nervous system pull on the heart at once. The sympathetic branch, the accelerator, speeds it up. The parasympathetic branch, mostly the vagus nerve, slows it down and does so quickly.

When you are rested, the vagus nerve is active and the gaps between beats vary noticeably. When you are stressed, ill, short of sleep or still processing last night's drinks, the accelerator dominates, the vagus pulls back, and the beats become metronomic.

So a high HRV generally means a nervous system with slack in it, and a low HRV means one running the engine. That is the whole concept.

You can feel a version of it. Breathe in slowly and the heart speeds slightly; breathe out and it slows. That rhythm is vagal, and it is where a good part of the measured number comes from.

The unit, and why brands disagree#

The most common metric is RMSSD, in milliseconds, which measures beat-to-beat differences. Some platforms report SDNN instead, which captures longer-term variation and gives larger numbers. Others report a proprietary score out of 100 and never tell you which underlying metric they used.

These are not interchangeable. An RMSSD of 45 and an SDNN of 45 are different states.

On top of that, devices sample at different times, for different lengths, using different filtering. An Apple Watch, an Oura ring, a Garmin and a chest strap will all produce different numbers from the same night.

The bands, with the caveat that matters more than the bands#

These are rough overnight RMSSD ranges in milliseconds in which most healthy adults fall. The spread inside each row is the point, not the row.

A healthy 45-year-old can sit at 22 or at 70. Both are normal for that person. That is why no medical guideline sets an HRV threshold, and why any product that tells you your HRV is "poor" is telling you about its own scoring, not about you.

Use your own seven-day rolling average. Everything useful here is a comparison with yourself.

What moves it, ranked by reliability#

Alcohol deserves its position at the top. It is the one input that produces an unmistakable effect in almost everybody, and it does so at quantities people think of as harmless and on nights they thought they slept well. For many people this is the first time a device has ever shown them a personal consequence of a drink, which is a more persuasive argument than any guideline.

How to read your own trend without losing your mind#

Four rules keep this useful.

Ignore single nights. Night-to-night noise is large. A number 20 percent below yesterday means nothing on its own.

Use a rolling average and a normal range. Most platforms draw one. Inside the band is inside the band; there is nothing to interpret.

Only act when the number and the feeling agree. Three days low plus heavy legs plus poor sleep is information. Three days low while you feel fine is usually measurement.

Decide what you will do before you look. A low reading should map to one action, normally an easier training day. If a low reading just produces worry, the metric is costing you more than it gives, and the right move is to stop looking at it.

The one use that earns its place#

For anyone training seriously, the combination of a falling HRV, a rising resting heart rate and worse sleep, for several days, is a good early warning of under-recovery or an infection starting. It often appears a day before symptoms do.

That is a real and practical signal. It is also nearly the only one. Outside sport, the honest use of HRV is as a nudge about alcohol and sleep, and it is quite good at that.

What I get asked#

The question is nearly always "my HRV is 28, is that bad?"

It is not answerable, and that is the answer. Without knowing what their number was last month, 28 carries no information. The people who get value from HRV are the ones who watched their own band for a month first and now know what a real deviation looks like for them.

The pattern I am most careful about is the opposite of worry. Someone with genuine symptoms, breathlessness on the stairs or chest tightness, who is reassured because the watch says recovery is good. A device measuring beat timing knows nothing about a narrowed coronary artery. No wearable number should ever be the reason a symptom goes unexamined.

When a rhythm finding is a medical matter#

HRV itself is not a diagnostic test. Some related watch findings are:

• A repeated irregular rhythm notification, or a possible atrial fibrillation alert, needs a proper ECG. Watch alerts do produce false positives, and they also find real atrial fibrillation, which is treatable and matters.
• Palpitations with chest pain, breathlessness or fainting is an emergency, whatever the watch says.
• A resting heart rate that has risen by more than 10 beats and stayed there for weeks, with no new training load, is worth investigating. Thyroid, anaemia and infection all do this.
• Very low HRV with dizziness on standing, in someone with long-standing diabetes, can reflect autonomic neuropathy and is worth raising.

The bottom line#

HRV is a real physiological measure reported by consumer devices with far more precision than it deserves. The number means nothing next to anyone else's and a great deal next to your own average from last month. Watch the band rather than the night, act only when the number and how you feel point the same way, and accept its most valuable finding: that alcohol and short sleep cost more than you thought, measured on you rather than argued at you.