Three Apple Watch Numbers That Should Change How You Train (And One That Shouldn't)

A typical Apple Watch quietly captures over a dozen metrics every day. Heart rate, HRV, resting heart rate, sleep stages, blood oxygen, wrist temperature, VO₂ max estimate, walking heart rate average, six different ring-closure trends. The Health app stacks them all in front of you with the same visual weight, as if every line on every chart deserved equal attention.

It doesn't. For someone who lifts and wants the data to actually inform a training decision, three of those numbers carry most of the signal, and one of the loudest ones carries almost none.

This is what a few years of looking at Apple Health data alongside actual session logs has taught a small team building a training app. Some of it is supported by research that's easy to find. Some of it is the kind of thing you only notice when you compare the dashboard to how the workout actually felt.

The First Number: HRV (And Why Yours Probably Reads Wrong)

Heart rate variability is the small, beat-to-beat fluctuation in the timing of your heartbeats. A relaxed, well-recovered nervous system produces more variability. A stressed, fatigued, or fighting-something one produces less. Apple measures it overnight, usually during sleep, and reports an SDNN value in milliseconds.

The first thing most people do with their HRV number is ask whether it's "good." That's the wrong question. Healthy adults span roughly 20 to 75 ms depending on age, sex, fitness, posture during measurement, and a dozen other things. Comparing your number to a population chart is about as useful as comparing your bodyweight without your height.

The number that matters is your own seven-day rolling average and how today sits against it. Plews and Laursen, who basically wrote the book on HRV-guided endurance training, have argued for years that the rolling average filters out daily noise and surfaces the trend that actually predicts performance. Vesterinen and colleagues ran a controlled study in 2016 where runners who let HRV trends decide their hard days outperformed runners on a fixed plan, even though the HRV group ended up training slightly less.

What does that mean on a Tuesday morning?

If your seven-day rolling HRV is 48 ms and you wake up at 46, that's noise. Train. If you wake up at 32, something is loud — poor sleep, illness coming, hard session two days ago, three drinks last night, a fight with your partner. The body does not separate those. It just registers stress.

The mistake people make is treating one bad HRV reading as a verdict. A single dip means the past 24 hours were rough. Two consecutive dips below baseline is when the signal becomes worth acting on. Three is when you should probably stop trying to PR.

The Second Number: Resting Heart Rate

If HRV is the high-resolution camera, resting heart rate is the low-resolution one with better stabilization. It moves more slowly. It's harder to spike with one bad night of sleep. And precisely because it's slower, when it does move, it usually means something.

Apple computes RHR using your lowest sustained heart rate during the day, typically caught while you're sitting still or sleeping. For most lifters, the absolute number lands somewhere between 48 and 72 bpm depending on cardiovascular conditioning, with trained athletes often dropping into the low 40s.

The signal isn't the number. It's the drift.

Endurance coaches have used a 5-to-7 bpm sustained increase over your normal RHR as a warning flag for over-reaching since at least the 1990s. The reasoning is mechanical: when the body is fighting accumulated load, an infection, dehydration, or inadequate sleep, it raises baseline cardiac output to compensate, and that shows up as a steady creep in RHR over several days.

In practice this looks like a chart that's been hovering around 54 for two months and slowly climbs to 60 over the course of a week. The HRV chart for the same week might be jumpy and inconsistent, which is normal — but the RHR creep is unambiguous. If you see both at the same time, that's not a coincidence. That's the body telling you it's running on a smaller fuel tank than it was a week ago.

The corrective action isn't dramatic. It's usually one fewer hard session this week, two if the trend continues, and an honest look at sleep and stress.

The Third Number: Sleep, But Not the Way You Think

Sleep is where most fitness writing gets sloppy. The standard headline is "you need eight hours" and the standard advice is "go to bed earlier." Neither of those engages with what the data actually shows about training.

The cleanest research on this comes from Cheri Mah's group at Stanford, which extended sleep in college basketball players from their habitual ~6.5 hours to a target of 10 hours in bed for several weeks. Free-throw accuracy went up. Sprint times went down. Reaction times improved. The effect size was large enough that the study has been cited a thousand times, including in places that misread it.

Here's the part most summaries skip: the gains showed up after weeks of extended sleep, not after one night. And the deficits — when athletes were short on sleep — accumulated across multiple nights before showing in performance.

For a lifter, this means a few specific things.

A single night of five hours probably won't tank a session if your week was otherwise solid. The body has reserves for this. Try to PR? Probably not the day. Hit a planned session at 80% load? Usually fine.

Three nights of five-to-six hours in a row is a different story. By the third morning, perceived exertion is up, top-end strength is down 3 to 5%, and the session that felt like an 8/10 RPE last week feels like a 9/10 today even though the weight is the same. This is when the smart move is to drop the top set or compress the session, not push through.

Apple's Sleep Score in watchOS 11+ tries to compress duration, consistency, and sleep stages into a single 0-100 number. It's not a bad heuristic, but it's a heuristic. Look at the duration trend across your last three to five nights and trust that more than the score itself.

The One to Ignore: Activity Rings

This is going to sound contrarian. The activity rings are the most-marketed, most-shared feature of the entire Apple Watch ecosystem, and for someone trying to train seriously, they're close to useless.

The Move ring tracks active calorie burn. The Exercise ring tracks minutes above a moderate heart rate threshold. The Stand ring tracks whether you stood up at least once per hour for 12 different hours. None of those three is closely related to whether your training is actually producing adaptation.

Active calories is a vanity metric that confuses output with input. Burning 800 active calories on a Tuesday tells you nothing about whether you got stronger, whether your sleep tonight will be good, or whether tomorrow's session should change. Two lifters can do the same productive 45-minute strength session and post wildly different active calorie numbers depending on body size, hydration, and how the Watch's algorithm felt that morning. Neither training adapted differently because of the calorie count.

The Stand goal is even less connected to training. Twelve isolated minutes of standing across a 12-hour window measures fidget frequency, not metabolic load. It's a behavioral nudge for sedentary office workers, repackaged as a fitness goal.

The Exercise minutes ring at least correlates with effort, but it correlates equally well whether you spent 30 minutes doing junk volume in a slow circuit class or 30 productive minutes under a barbell. Both close the ring. Only one moves you forward.

The bigger problem with rings is what they do to attention. The dashboard rewards you for closure regardless of quality. Over time, people optimize for the metric they can see — and what they can see is rings, not adaptation. They take an extra walk to close Move, skip a planned rest day to keep an Exercise streak alive, and end up doing slightly more low-quality activity in exchange for slightly worse recovery for the high-quality work.

Stop closing the rings. Start watching the three signals that actually move when your training is on or off track.

Putting It Together

If your goal is "look at my Watch in the morning and decide what kind of session today should be," the heuristic is simpler than the dashboard makes it look.

Start with the seven-day rolling HRV. If today is within roughly 10% of that average, training as planned is the default. If today is meaningfully below — say, more than 15% under baseline — and yesterday was also below, the smart move is to lower top-end intensity and trim a working set or two. RPE 7 instead of RPE 9. Three working sets instead of four.

Cross-reference with RHR. If RHR is also creeping up over the last several days, that's a stronger signal than HRV alone. The body is telling you the same thing through two different channels.

Check the trailing three nights of sleep duration. Two or more nights below your normal range is the threshold where it starts mattering for top-end strength. Below that, treat sleep as background context and let HRV plus RHR drive the call.

Ignore the rings while you're doing this. They're not telling you anything about whether to train.

The Apple Watch is, for what it costs, a remarkable piece of physiological measurement hardware. It just happens to ship with a software layer optimized for engagement rather than training decisions. The good signal is in there. It's mostly a question of which charts you choose to look at.

Sources Worth Reading

For anyone who wants to go deeper on the actual research, three places are worth the time.

Plews and Laursen's review work on HRV-guided endurance training is the cleanest practical synthesis of how to use rolling averages for training decisions. Vesterinen et al. (2016) is the controlled trial showing the approach actually outperforms a fixed plan. Mah et al.'s sleep extension study at Stanford remains the cleanest demonstration that sleep is a ceiling on athletic output, not just a recovery nice-to-have.

None of that research was done with an Apple Watch in mind. The Watch is just one of the more accessible ways to surface the signals those studies were measuring with chest straps and lab equipment. The numbers are the same. The interpretation is the same. The dashboard is louder.