Garmin Metrics Explained: What Every Number On Your Watch Actually Means
Compartilhar
A modern Garmin shows you about twenty numbers and explains almost none of them. Worse, it gives them all the same confidence — a figure measured off your skin sits next to a figure extrapolated from a guess about your age, and the two look equally authoritative.
This page is the index. Every metric, what it actually is, and the part nobody else prints: how much it deserves to be trusted.
The organising idea is simple. Stop asking what a number means and start asking what it is for. There are only four questions a training metric can usefully answer:
- How hard was that session?
- Am I recovered?
- Am I getting fitter?
- Is something wrong?
Every number on your watch belongs to one of those four. Once you know which, you know when to look at it — and when to ignore it.
The whole lot, on one screen
| Metric | What it is for | What it actually is | How much to trust it |
|---|---|---|---|
| Training effect (aerobic) | How hard was that? | A 0.0–5.0 score for one session's effect on aerobic fitness | Medium — scaled against your max heart rate |
| Training effect (anaerobic) | How hard was that? | The same scale for the short, above-threshold part | Medium — same dependency |
| Training load (per session) | How hard was that? | One figure for the cost of one activity | Medium — inherits the same error |
| Acute load / 7-day load | How hard was that? | Recent load, weighted towards the last few days | Medium — directionally useful |
| Load ratio | How hard was that? | Recent load against your longer-term baseline | Medium — unreliable after a layoff |
| Load focus | How hard was that? | How your load splits across low aerobic, high aerobic, anaerobic | Medium-high — the most useful load screen |
| Intensity minutes | How hard was that? | Weekly moderate and vigorous minutes vs a health guideline | Medium — a health metric, not a training one |
| Body Battery | Am I recovered? | A 0–100 energy estimate, charged by rest, drained by effort | Medium-high as a trend, not as an absolute |
| HRV status | Am I recovered? | Overnight heart rate variability vs your own baseline | High — once you have weeks of data |
| Sleep score | Am I recovered? | A 0–100 score from duration, stages and overnight physiology | Mixed — duration decent, stages weak |
| Stress score | Am I recovered? | A 0–100 daytime figure from waking heart rate variability | Medium — cannot separate work stress from training |
| Training readiness | Am I recovered? | Six recovery inputs compressed into one 0–100 score | Medium — inherits every input's error |
| Recovery time | Am I recovered? | Hours until the watch thinks you can train hard again | Low-medium — the most frequently wrong number here |
| VO2 max | Am I getting fitter? | An estimate of maximal oxygen uptake from pace vs heart rate | Medium as your own trend, poor as an absolute |
| Fitness age | Am I getting fitter? | Your VO2 max estimate restated as an age | Same as VO2 max — it is mostly the same number |
| Training status | Am I getting fitter? | A label — Productive, Unproductive — from load plus fitness trend | Medium — opaque logic, slow to settle |
| Endurance score | Am I getting fitter? | A score for sustaining prolonged effort over hours | Unproven — newest and least validated |
| Race predictor | Am I getting fitter? | Estimated finish times for 5K, 10K, half marathon and marathon | Low-medium — commonly optimistic over long distances |
| Resting heart rate | Is something wrong? | Your lowest sustained heart rate, usually overnight | High — close to a direct measurement |
| Respiration rate | Is something wrong? | Breaths per minute, derived from the heart rate signal | Medium — fine as a trend, noisy night to night |
| Pulse Ox (SpO2) | Is something wrong? | Estimated blood oxygen saturation, read at the wrist | Low — noisy, and not a medical device |
The dividing line nobody tells you about
Here is the split that explains most "my Garmin data makes no sense" complaints, and it is not the split the menus imply.
About half of these numbers are built on top of your maximum heart rate — a value that, unless you changed it yourself, your watch guessed from 220 minus your age. Training effect, training load, acute load, load focus, intensity minutes, recovery time, VO2 max and training status all scale intensity as a percentage of that ceiling.
The other half do not care what is in your profile. HRV status, resting heart rate, respiration rate, Pulse Ox, sleep duration and Body Battery come out of sensor data. They have their own problems — a wrist optical sensor is not a laboratory instrument — but they are not corrupted by a settings mistake.
That is why you can see a perfectly sensible HRV status and a wildly implausible recovery time on the same morning and conclude the watch is broken. It is not. One of those numbers was measured; the other was calculated from an assumption about everyone born in your year.
So if your intensity-derived numbers consistently feel wrong, fix your max heart rate before troubleshooting anything else. It is a five-minute change that corrects eight metrics at once. Your heart rate zones sit downstream of the same assumption, and there are four separate reasons they can be wrong — a wrong ceiling is only the first of them.
Group 1: How hard was that session?
Training effect — aerobic and anaerobic
Two scores from 0.0 to 5.0 at the end of every activity. Aerobic training effect rates what the session did for your endurance base; anaerobic training effect rates the short, hard, above-threshold portion. The bands run from no benefit at the bottom up to overreaching at the top. These are the most immediately useful numbers your watch produces, because they tell you whether the session you intended was the session you did. An easy run that comes back with a high aerobic effect was not easy. How the two scores work, and what to do when they contradict your plan.
Training load, acute load and load ratio
Training load is one number for the cost of one activity. Acute load — "7-day training load" on older watches — accumulates those into a rolling figure weighted towards recent days. Newer watches add a load ratio comparing that against your longer-term baseline.
Read it as a trend, never as a target. It reliably misleads in one place: coming back from a break, a couple of modest sessions against a decayed baseline can spike the ratio and trigger warnings. That is an artefact of dividing two numbers, not evidence you overdid it. Which of the three load numbers you are actually reading, and why load focus is the one worth your attention.
Load focus
The most underrated screen in Garmin Connect. It splits recent load into low aerobic, high aerobic and anaerobic buckets, each against a target range. If you have wondered whether your training is genuinely polarised or just permanently moderate, this screen answers it — and most recreational runners are shocked by how little real low-intensity work they do.
Intensity minutes
Weekly minutes of moderate and vigorous activity, with vigorous minutes counting double. A public-health metric, not a training one: it tracks you against activity guidelines rather than against your race. A long easy ride banks an enormous number of intensity minutes while changing very little about your fitness; a sharp interval session banks fewer.
Group 2: Am I recovered?
Body Battery
A 0–100 gauge that charges while you rest and drains while you work, stress or stay awake, driven mainly by heart rate variability plus activity and sleep. Its honest use is as a trend — the absolute figure moves for reasons that have nothing to do with training, so "mine is 42" means much less than "mine has not got above 60 in a fortnight". What drains it, why it sometimes refuses to charge overnight, and what a realistic morning number looks like.
HRV status
Your overnight heart rate variability, averaged over a week and compared against the baseline your own watch built for you. It is the best recovery signal a wrist device produces and the one needing most patience: the baseline takes weeks of consistent overnight wear before the status means anything.
Two common mistakes. HRV is only meaningful against your own history, so comparing your figure with someone else's tells you nothing. And a status can be flagged unbalanced for sitting above your usual range as well as below it. What each status means and how long it takes to clear — and separately, why the "good HRV by age" tables are more misleading than helpful.
Sleep score
A 0–100 score for last night from duration, stage estimates and overnight physiology. Trust the duration more than the breakdown: estimating how long someone slept from movement and heart rate is tractable, while classifying a given ten minutes as deep or light from the same signal is much harder. The stage graph is the part most likely to disagree with your memory of the night. What goes into the score, and why a bad night sometimes scores well.
Stress score
A daytime 0–100 figure from heart rate variability while you are awake. Its crucial limitation is that it has no idea why your nervous system is busy — a hard week at work and a hard week of intervals look alike to it. That makes it useful, because total load is total load, but a high score is not automatically a training problem. How it is measured, and when a high score is meaningless.
Training readiness
Six recovery inputs — last night's sleep, sleep history, HRV status, recovery time, acute load and stress history — compressed into one score out of 100. Compression destroys information, which is why "my readiness is 34" is never a complete answer. Open the sub-scores and find which of the six is dragging. Here is how, input by input.
Recovery time
A countdown in hours until your watch thinks you can train hard again. It is the metric people complain about most, and the complaints are usually justified: it is driven by time spent at high intensity, and intensity is a percentage of a max heart rate the watch probably guessed. Set that ceiling too low and every session buys a punitive recovery time forever. The four things that inflate it, and when to override it.
Group 3: Am I getting fitter?
VO2 max and fitness age
VO2 max is the maximum rate at which your body can take in and use oxygen. Your Garmin does not measure it — it infers it from the relationship between your pace and your heart rate on qualifying outdoor runs, so the estimate is only as good as its inputs. Fitness age is the same number wearing a friendlier hat: on current models it comes largely from that estimate set against population norms for your age and sex, so it carries every error the estimate carries. How accurate it really is, what a good value looks like, and why a falling VO2 max is usually a heart rate data problem rather than a fitness problem.
Training status
A one-word verdict — Productive, Maintaining, Unproductive, Peaking, Recovery, Detraining, Overreaching, and on some models Strained — from your load trend set against your fitness trend. It is the closest thing your watch has to coaching, and it is slow, opaque and easily confused. "Unproductive" frequently means "your VO2 max estimate drifted down while your load went up", which can be a data artefact rather than a training mistake. What each status is reacting to, and what to actually change.
Endurance score
A newer, higher-end metric scoring your capacity for prolonged effort rather than your peak capacity, built from longer-term training history. Conceptually welcome — VO2 max says nothing about holding a pace for three hours. Treat it as the least proven number on your watch: recent, on fewer models, and with far less independent evidence about what it tracks than the older metrics. What actually moves it, and why a good VO2 max estimate can sit beside a poor endurance score.
Race predictor
Predicted finish times for 5K, 10K, half marathon and marathon, leaning on your VO2 max estimate alongside your training history. The reported pattern is consistent: sensible at 5K and 10K, optimistic as the distance grows, because a VO2-max-driven model knows about your engine and much less about whether you have ever run for three hours. If the marathon prediction looks flattering, it probably is. How much to trust it at each distance, and why the error only ever goes one way.
Group 4: Is something wrong?
Resting heart rate
Your lowest sustained heart rate, normally captured overnight, and one of the most trustworthy numbers on your watch precisely because it is close to a direct measurement rather than a model. Its value is in the trend: a resting heart rate several beats above your normal for a few consecutive days is a well-established early flag for illness or accumulated fatigue, often before you feel anything. Read it alongside HRV status and the other signs of over-training. A rise has about seven plausible causes and only two of them need you to change the training — how to tell them apart.
Respiration rate
Breaths per minute, derived from your heart rate signal rather than measured directly. Night-to-night variation is normal, but a sustained overnight rise — especially alongside a raised resting heart rate — is another early illness flag. What moves it overnight, and why the figure during exercise means almost nothing.
Pulse Ox (SpO2)
An estimate of blood oxygen saturation from how light is absorbed at your wrist. Two honest caveats: wrist readings are much noisier than a fingertip clinical device, and Garmin itself states these watches are not medical devices. Its legitimate uses are narrow — tracking acclimatisation at altitude, and spotting overnight desaturation worth raising with a doctor. Nothing on this page is medical advice. If a reading worries you, or you have symptoms, that is a conversation for a clinician, not a watch. How accurate wrist SpO2 really is, when it is worth switching on, and what all-day tracking costs your battery.
Which one should you look at this morning?
Twenty metrics is not a morning routine. A short answer:
- Deciding today's session: training readiness — or HRV status plus Body Battery if your watch lacks it. One glance.
- Checking a session went to plan: aerobic and anaerobic training effect, right after you finish.
- Checking the month went to plan: load focus and training status. Both are weekly-or-slower metrics; looking at them daily tells you nothing.
- Worried you are getting ill: resting heart rate and respiration rate trends, read together.
- Everything looks wrong at once: your max heart rate. See the dividing line above.
One caveat that saves a lot of anxiety: these metrics read very differently if you are not a pure runner. Garmin's fitness model is built around the pace-to-heart-rate relationship in steady running. Sled pushes, wall balls and heavy carries produce enormous real physiological cost that the model either ignores or mis-scores, so a hybrid athlete can train extremely hard and watch their training status say "Unproductive". Worth understanding if you race Hyrox or anything like it.
The thing all of these have in common
Your watch is excellent at measuring and poor at telling you what to do next. Every number above is an input, not an instruction. Recovery time does not know you have a race in nine days. Training status does not know you deliberately took an easy fortnight. Readiness does not know the reason you slept badly is a new baby rather than over-reaching.
There is also a circularity at the centre of the system: the intensity-based metrics are all percentages of a maximum heart rate that was estimated. The only way out is to anchor on something measurable. Your lactate threshold is that thing — the intensity where lactate starts accumulating faster than you clear it, a real physiological boundary rather than a fraction of a guess. Your watch tries to estimate it too, though its automatic detection is fussy and fails silently.
Turning twenty inputs into one decision is the actual work, and it is the gap between what a watch does and what a coach does. If you want that done for you, Lactate Training reads your synced Garmin activities and health data and builds and adapts a plan from them, with zones anchored to your lactate threshold rather than to 220 minus your age. Whether software can really do a coach's job is worth reading about honestly: where an AI coach beats a human one and where it does not, and what it means to put your own training data in front of a general-purpose model like ChatGPT or Claude. Before you look anywhere else, it is worth knowing what the coaching already built into Garmin Connect does well and where it stops.
Frequently asked questions
Which Garmin metrics are measured and which are estimated?
Resting heart rate, respiration rate, heart rate variability and sleep duration come close to direct measurement from sensor data. VO2 max, training effect, training load, recovery time, training status, race predictions and fitness age are all models — calculated from your heart rate expressed as a percentage of a maximum heart rate your watch most likely guessed from your age.
Which Garmin metric should I check in the morning?
One: training readiness, if your watch has it, because it already bundles sleep, HRV, recovery time, load and stress. If your watch does not have readiness, use HRV status together with Body Battery. Checking all twenty metrics every morning produces anxiety rather than information.
Why do my Garmin metrics contradict each other?
Usually because some are built on your max heart rate setting and some are not. A plausible HRV status next to an absurd recovery time is the classic signature: one was measured, the other was calculated from a guessed ceiling. Fix the max heart rate and the contradiction normally resolves.
Why did all my Garmin numbers change at once?
Something upstream changed. Editing your user profile, a firmware update, switching between optical heart rate and a chest strap, or a new max heart rate value will all shift every intensity-derived metric together. Rolling metrics such as VO2 max, training status and acute load are built from trailing windows, so expect a week or two of instability before they settle.
Can I trust Garmin's VO2 max and race predictions?
As a trend for yourself, reasonably. As an absolute number, or for comparing yourself with anyone else, not really. Both are estimates derived from your pace against your heart rate, and race predictions in particular tend to be optimistic over longer distances because the model knows about your aerobic engine and much less about your endurance.
Do I need a chest strap for Garmin metrics to work?
Not for most of them, but accuracy improves markedly with one and a few features effectively require it. Wrist optical heart rate struggles during intervals and with wrist movement, and because so many metrics are intensity-derived, bad heart rate data propagates into all of them. Garmin's lactate threshold detection in particular will not run without a compatible strap.
Twenty numbers, one decision
Reading the whole dashboard is a skill. Not having to read it is better. Lactate Training pulls your Garmin activities and health data in, builds a plan from them and adapts it as they move, in 24 languages.