The Heart Screen: Four Signals, and What Each Is For

Resting heart rate

Read from your platform, which computes it by sampling through the day and taking the lowest sustained values from periods when you were genuinely still, usually overnight.

It contributes 30% of recovery score, inverted — lower is better, so the z-score is computed as baseline minus today rather than the reverse.

The single most actionable pattern: a rise of five or more beats per minute over your baseline is usually the first sign of illness, and it typically shows up a day before you feel anything. That is a genuinely useful early warning and it is the reason to look at this screen at all.

It also rises with accumulated training fatigue, poor sleep, alcohol, heat and dehydration. Which is why it is read as a deviation alongside other signals rather than as a verdict on its own.

Apple often writes this hours late, so an early-morning glance can show yesterday's figure. The screen labels which day a reading came from rather than presenting a stale number as current.

Walking heart rate average

The average heart rate measured while you walk steadily on flat ground — not running, not climbing stairs, not carrying anything.

This is the closest thing on the screen to a fitness measure, and it is an underrated one. The same walk gets metabolically cheaper as your aerobic base improves, so this number drifts down over months of consistent cardio and drifts back up when you stop.

Two properties make it better than it looks:

It is more stable than resting heart rate. Less sensitive to one bad night, one late meal, one warm room.

It is submaximal and repeatable, which means it is measuring something about your capacity rather than something about your acute state.

It is frequently missing, because platforms only write it after enough sustained walking. On a day spent mostly sitting there is nothing to compute, and the screen says "none this day" rather than showing a zero.

Heart rate variability

The largest single input to recovery score at 50% of the mix. Base reads the SDNN value your watch computes overnight rather than deriving its own.

Meaningful only against your own baseline. HRV varies enormously between people for reasons of age and genetics that no training changes, so a cross-person comparison ranks demographics rather than readiness. There is no percentile here and there never will be.

It also has a specific limitation worth knowing: HRV responds to metabolic and autonomic stress more than to mechanical stress. A heavy squat session produces a large amount of muscle damage and neural fatigue and a relatively modest HRV response — so it under-reads lifting load, which is part of why the deload trigger also reads RIR drift. The full picture.

Respiratory rate

Read from the health store as the average of the samples your watch took across the day.

Base does not compute it and does not feed it into any score. It is shown for one reason: a rise of about two breaths per minute over your baseline is an early illness signal that none of the other numbers catch as early.

It has no weighting in recovery score because there is no defensible weighting to give it — the evidence for its use as a training readiness input is much thinner than for HRV or resting heart rate. Including it in a score would be inventing a weight. Showing it and saying what it commonly indicates is the honest treatment.

Zones, and the refusal to guess your age

Heart rate zones are percentages of a maximum, and maximum is estimated from your age.

With no birth year on file there is no honest maximum to take a percentage of, so no zones are shown at all. Not a default, not a population average, not an assumed forty. Nothing.

That is the same rule that governs every missing input in the app, and it costs something — a user without a birth year gets a less useful screen. The alternative costs more: a set of zone boundaries computed from an assumption, presented identically to ones computed from your actual age.

Where zones are available, they are used for the easy and long endurance bands specifically: 60 to 70% of maximum. On an easy day, obey the heart-rate band rather than pace — pace drifts several percent with heat and hills while heart rate reflects the actual physiological cost. On interval days the reverse holds, because heart rate lags the effort by thirty to sixty seconds. How the bands work.

Reading it as a hybrid athlete

The systemic signals here do not know which discipline caused the fatigue, and that is exactly right.

A lifting-only fatigue model misses the athlete whose problem is a spiked endurance week. An endurance-only model misses the one whose problem is three heavy lower-body sessions. Resting heart rate and HRV catch both — and a hybrid athlete's actual problem is almost always that the two loads combined exceeded what either alone would have.

What this screen will not do

It will not diagnose anything. A rising resting heart rate is a common early illness signal. It is also many other things, and Base is a training log rather than a medical device. Persistent abnormalities are a doctor's question.

It will not rank you. No percentiles, on any of the four.

It will not show a number it had to guess. No data means "no data for this day".

What to do next

The recovery screen is where three of these four become a decision about today's session, and the HRV research covers what moves the largest of them that has nothing to do with training.