HRV and Recovery: What the Gaps Between Your Heartbeats Actually Tell You

Your heart doesn't beat like a metronome, and that's the whole point. The tiny, constantly-shifting gaps between one beat and the next are one of the best physiological windows you have into how recovered you actually are, if you know how to read them.

What HRV actually measures

Heart rate variability (HRV) is exactly what it sounds like: the variation in time between consecutive heartbeats. The gap between beat one and beat two is never precisely the gap between beat two and beat three. That jitter looks like noise, but it isn't: it's a live readout of your autonomic nervous system, the automatic control layer that runs your heart, lungs, and stress response without you thinking about it.

That system has two branches pulling in opposite directions. The sympathetic branch is your "fight-or-flight" accelerator: it fires when you're stressed, under-slept, fighting an infection, or hammered by a hard training block. The parasympathetic branch is your "rest-and-digest" brake: it dominates when you're calm and recovered. When the parasympathetic side is in charge, your heart rhythm has more variability. When you're stressed or fatigued, the rhythm gets more rigid and metronomic, and HRV drops.

So the headline is simple: higher HRV generally signals better recovery. But there's a giant asterisk on the word "higher," and getting that asterisk wrong is where most apps quietly fall apart.

Higher than your own baseline, not higher than anyone else's

Here's the part most tools get wrong. HRV varies enormously between people. Genetics, age, and fitness mean one person can sit comfortably at 90ms while another well-trained, perfectly healthy person lives at 35ms. Comparing your number to a population chart tells you almost nothing. A reading that's a red flag for one person is a Tuesday for another.

What actually carries information is the deviation from your own normal. This is why RepTrack (like Whoop, Oura, and Garmin before it) scores HRV relative to your personal baseline, not against a population table (Whoop developer docs; Oura Readiness docs; Garmin HRV Status manual). Specifically, RepTrack builds a rolling 30-day baseline from your own history using robust median and MAD (median absolute deviation) statistics, which shrug off the occasional weird reading instead of letting it drag the whole average around. A score of 70 means "exactly at your personal baseline." Above 70 means today's signals are stronger than your recent norm; below means they're softer.

That baseline needs data to exist. Until RepTrack has at least seven days of readings, it can't trust a personal baseline yet, so it shows a clearly-labeled "calibrating" state and leans on population reference values in the meantime. It won't pretend to know your normal before it does.

SDNN vs RMSSD, briefly

You'll see two HRV metrics thrown around. Both are time-domain measures: they work directly with the intervals between beats.

• SDNN is the standard deviation of all those beat-to-beat intervals over a window. It's a broad measure of total variability. Apple Watch reports HRV as SDNN, sampled periodically throughout the day and, importantly, during sleep.
• RMSSD is the root mean square of the successive differences between beats. It's more sensitive to that parasympathetic "brake," which makes it popular for short morning readings.

You don't need to memorize the difference. The useful takeaway: these time-domain metrics (SDNN and RMSSD) are the robust part of HRV: the piece that holds up when validated against a medical-grade ECG (Dial et al. 2025). Fancier frequency-domain breakdowns are far more sensitive to measurement conditions, so RepTrack builds on the sturdy stuff.

Why HRV gets log-transformed

HRV is log-normal: its distribution is skewed, with a long tail of high readings. If you average raw values, a few big numbers yank the mean upward and distort everything. So before running any statistics, RepTrack (and the research it's built on) applies a natural-log transform to HRV, which straightens the distribution out so the math behaves (Plews et al. 2013; Shaffer & Ginsberg 2017). It's a quiet technical step, but it's the difference between a stable trend line and a jumpy, misleading one.

One reading tells you almost nothing

If you take a single HRV measurement this morning and panic because it's low, stop. A lone reading is noisy. Your position, the time of day, last night's drink, a stressful email, or the flight of stairs you just climbed can all swing one number. The signal lives in the trend over days and weeks, not in any single data point (Plews et al. 2013). This is precisely why RepTrack scores against a 30-day rolling baseline instead of reacting to today in isolation. A single low morning after a hard session is expected. A steady multi-day slide is the message.

Why sleep readings beat daytime spot checks

The cleanest HRV data comes when you're asleep. You're lying still, in a consistent position, hours removed from caffeine, movement, and the day's stress spikes. Those controlled conditions make sleep-time readings far more comparable night to night.

That's why RepTrack prefers HRV captured during sleep and weights it accordingly. HRV is the single largest input to the recovery score, and the score is most trustworthy when that HRV came from sleep. When only daytime spot-checks are available, RepTrack doesn't pretend they're as good. It still uses them, but it lowers the confidence it displays alongside your score, so you know exactly how much signal is standing behind the number.

The honest caveat: not every wearable even gives you HRV

This is the caveat most recovery apps bury, and RepTrack puts it front and center: some wearables don't export HRV to Apple Health at all.

• Garmin syncs plenty of metrics to Apple Health, but none of them is HRV.
• Oura exports sleep, heart rate, and respiratory rate, but not HRV or wrist temperature.

RepTrack can't invent data it was never given. So instead of fabricating an HRV-weighted score for those users, it builds an honest one from what it does have (resting heart rate and sleep) and renormalizes the math so the available signals carry the full weight. The result is a real recovery score, genuinely useful, just carried at a lower, clearly-displayed confidence. An Apple Watch delivers the full picture; a Garmin or Oura delivers an accurate subset; an iPhone alone gets no physiological score at all, just a clearly-labeled "Training Readiness" estimate. A day with no signal shows a gap, never a fake number. That honesty contract is the whole design.

For the full breakdown of how these signals combine into a single recovery score (the weights, the bands, and why sleep is scored differently from everything else), see the flagship, "How Your Recovery Score Actually Works."

Practical levers that actually move your HRV

You can influence HRV. The big ones, roughly in order of impact:

• Sleep. The strongest lever by far. Duration and quality both feed HRV, and adults generally need about 7–9 hours (Hirshkowitz et al. 2015).
• Alcohol. Even a couple of drinks reliably crushes HRV for that night. One of the most consistent effects you'll see in your own data.
• Stress load. Psychological and physiological stress both suppress parasympathetic tone.
• Training load. Hard blocks push HRV down; that's expected and even desirable in the short term. The popular acute-to-chronic workload ratio has a rough sweet spot, but treat it as a contested heuristic, not a law (Gabbett 2016 and its critics).
• Measurement consistency. Measure the same way, ideally during sleep, every night. Half of a "crashing HRV trend" is often just inconsistent conditions.

HRV is a training signal, not a medical readout

One last thing, and it's not a footnote. HRV is a training-readiness signal: a tool to help you decide whether to push or pull back on a given day. It is not a diagnosis, a disease screen, or a medical device, and RepTrack never presents it as one. If a persistent change in your readings worries you, that's a conversation for a clinician, not an app. RepTrack's job is to read your own signals honestly, show you the trend and the confidence behind it, and never fabricate a number to fill a gap.

The bottom line

HRV is one of the clearest windows into recovery you have, but only when you read it as a trend against your own baseline, not a single number and never a population chart. Higher-than-your-normal generally means recovered; a multi-day slide means back off. Sleep-time readings are the gold standard, log-transformed trends beat raw spot-checks, and if your wearable can't hand over HRV, an honest resting-heart-rate-and-sleep score at lower confidence beats a fabricated one every time. Track the trend, keep your measurement consistent, and let the confidence label tell you how much to trust today's read.

Sources

• Plews DJ, Laursen PB, Stanley J, Kilding AE, Buchheit M. "Training adaptation and heart rate variability in elite endurance athletes." Sports Medicine, 2013.
• Shaffer F, Ginsberg JP. "An Overview of Heart Rate Variability Metrics and Norms." Frontiers in Public Health, 2017.
• Dial et al. "Validation of Apple Watch heart rate variability (SDNN) against ECG." 2025.
• Hirshkowitz M, et al. "National Sleep Foundation's sleep time duration recommendations." Sleep Health, 2015.
• Gabbett TJ. "The training-injury prevention paradox: should athletes be training smarter and harder?" British Journal of Sports Medicine, 2016 (and subsequent acute:chronic workload ratio critiques).
• Whoop Developer Documentation: recovery and HRV baseline scoring.
• Oura Readiness Score Documentation: personal-baseline scoring.
• Garmin HRV Status Owner's Manual: baseline-relative HRV status.