12 Wearables · 35 Factors

Compare recovery scores across 12 wearables.

See the 35 factors that move yours, and which scores are actually validated. They aren't interchangeable (two devices on one wrist disagree), and only 2 of 12 are validated.

Brands compared

Factors that move it

2 / 12

Scores validated

86%

Use HRV as an input

How each brand measures recovery

Every recovery score, side by side.

Two devices on one wrist will disagree: different inputs, baselines, and scales. Here's what each score actually reads, ranked by input coverage.

Full breakdown: WHOOP vs Oura vs Garmin compared

Input coverage

Signals fed into the recovery score, by brand

Same core ingredients: HRV (86% of scores), resting HR (79%), sleep (71%). The difference between brands is weighting, not raw signals.

12 brands · 7 signals

Wearable	HRV	Resting HR	Sleep	Body Temp	Resp. Rate	Training Load	Self-report	Signals
Apple Watch	6/ 7
RingConn	6/ 7
Amazfit / Zepp	5/ 7
Oura Ring	5/ 7
WHOOP	5/ 7
Polar	4/ 7
Samsung	4/ 7
Ultrahuman	4/ 7
Fitbit / Pixel	3/ 7
Garmin	3/ 7
Suunto	3/ 7
COROS	1/ 7

Get the Kygo app

What moves your score

35 factors that move your recovery score.

Anything that moves HRV, resting heart rate or sleep moves your score, always relative to your own baseline.

Read: what lowers your recovery score

Filter by category: 5 groups

Sleep

Nutrition

Training

Stress

Baseline

Baseline 4 factors

Baseline modifiers

4 factors · Shift your baseline

Baseline

Age

Shifts your baseline · High impact · Strong evidence

High

Baseline

Body composition (body fat %)

Shifts your baseline · Medium impact · Moderate evidence

Medium

Baseline

Menstrual cycle (luteal phase)

Shifts your baseline · Medium impact · Moderate–strong evidence

Medium

Baseline

Sex

Shifts your baseline · Medium impact · Moderate evidence

Medium

Get the free State of Wearable Accuracy report

Per-brand deep dive

HRV falls with age, so a "low" number for a 60-year-old can be perfectly normal. Compare to age norms and your own baseline, never to other people.

Effect on the signals: HRV declines with age (~80 ms in teens to ~25 ms over 75).
Evidence: Strong
Mechanism: Vagal tone and autonomic flexibility decline with age.
What to do: Judge your score against your own history and age-appropriate norms.

Source

Age, sex & fitness differences in HRV (PMC6686273)

Body fat percentage, not weight or BMI, is the informative measure: higher body fat tracks lower HRV and higher RHR, and it improves with fat loss.

Effect on the signals: Higher body fat tracks lower HRV and higher RHR; BMI alone is not independently associated.
Evidence: Moderate
Mechanism: Adipose tissue raises sympathetic tone and inflammatory load; fat loss restores autonomic balance.
What to do: Track body fat %, not the scale; sustainable fat loss lifts the baseline.

Body fat percentage & HRV sympathovagal balance (PMC3743431)

Normal physiology, not a problem: expect lower scores mid-to-late cycle as RHR and temperature rise and HRV dips.

Effect on the signals: RHR +~3–5 bpm (~3%), HRV down, temperature +~0.5°C in the luteal phase.
Evidence: Moderate–strong
Mechanism: Progesterone raises basal temperature and heart rate and lowers vagal tone in the luteal phase.
What to do: Expect and accept lower numbers in the luteal phase; compare within-cycle, not across phases.

Menstrual cycle & cardiac autonomic function / HRV (PMC5588411)

Women tend to run higher vagal tone, men higher sympathetic, and the gap narrows after about 55 (menopause).

Effect on the signals: Women: higher vagal tone; men: higher sympathetic. The gap narrows after ~55 (menopause).
Evidence: Moderate
Mechanism: Sex hormones modulate autonomic balance; the difference attenuates post-menopause.

Stress 4 factors

What helps

1 factor · Raises your score

Stress

Slow breathing / meditation

Raises your score · Medium impact · Moderate–strong evidence

Medium

What hurts

3 factors · Lowers your score

Stress

High altitude (acute)

Lowers your score · High impact · Strong at altitude evidence

High

Stress

Illness onset

Lowers your score · High impact · Strong evidence

High

Stress

Psychological stress

Lowers your score · Medium impact · Moderate–strong evidence

Medium

Get the free State of Wearable Accuracy report

Per-brand deep dive

Training 5 factors

What helps

4 factors · Raises your score

Training

Aerobic fitness (long-term)

Raises your score · High impact · Strong evidence

High

Training

Cold water immersion (post-exercise)

Raises your score · Medium impact · Moderate evidence

Medium

Training

Heat acclimation (repeated heat)

Raises your score · Medium impact · Moderate evidence

Medium

Training

Sauna (acute session)

Raises your score · Low impact · Mixed evidence

Low

What hurts

1 factor · Lowers your score

Training

Hard training / overreaching

Lowers your score · High impact · Strong evidence

High

Get the free State of Wearable Accuracy report

Per-brand deep dive

Nutrition 18 factors

What helps

11 factors · Raises your score

Supplement

Ashwagandha

Raises your score · Medium impact · Moderate evidence

Medium

Supplement

Glycine

Raises your score · Medium impact · Moderate evidence

Medium

Supplement

L-theanine

Raises your score · Medium impact · Moderate evidence

Medium

Supplement

Magnesium (bisglycinate)

Raises your score · Medium impact · Moderate–weak evidence

Medium

Supplement

Saffron

Raises your score · Medium impact · Moderate evidence

Medium

Supplement

Tart cherry (Montmorency)

Raises your score · Medium impact · Moderate (small pilots) evidence

Medium

Supplement

Chamomile extract

Raises your score · Low impact · Moderate evidence

Low

Supplement

Dietary nitrate / beetroot

Raises your score · Low impact · Moderate / mixed evidence

Low

Supplement

Melatonin

Raises your score · Low impact · Moderate (timing) evidence

Low

Supplement

Omega-3 (DHA)

Raises your score · Low impact · Moderate (RHR) evidence

Low

Supplement

Weak / don’t rely on (valerian, GABA, CBD, creatine)

Raises your score · Low impact · Weak evidence

Low

What hurts

7 factors · Lowers your score

Substances

Alcohol

Lowers your score · High impact · Strong evidence

High

Substances

Nicotine

Lowers your score · High impact · Strong evidence

High

Nutrition

Dehydration

Lowers your score · Medium impact · Moderate evidence

Medium

Substances

Evening caffeine

Lowers your score · Medium impact · Moderate evidence

Medium

Nutrition

High-glycemic meal / glucose spike

Lowers your score · Medium impact · Moderate (direction) evidence

Medium

Nutrition

Late or heavy meal near bed

Lowers your score · Medium impact · Moderate evidence

Medium

Substances

THC / cannabis

Lowers your score · Low impact · Minor evidence

Low

Get the free State of Wearable Accuracy report

Per-brand deep dive

Sleep 4 factors

What helps

1 factor · Raises your score

Sleep

Adequate, consistent sleep

Raises your score · High impact · Strong evidence

High

What hurts

3 factors · Lowers your score

Sleep

Sleep loss / sleep debt

Lowers your score · High impact · Strong evidence

High

Sleep

Circadian disruption (jet lag, shift work)

Lowers your score · Medium impact · Moderate evidence

Medium

Sleep

Hot bedroom / acute heat

Lowers your score · Medium impact · Moderate evidence

Medium

Get the free State of Wearable Accuracy report

Per-brand deep dive

Browse by category 35 total

Pick a category to explore its factors.

Sleep

Nutrition

Training

Stress

Baseline

Get the free State of Wearable Accuracy report

Per-brand deep dive

How each recovery score actually works.

Tap any brand for how its score is built, its scale and inputs, free vs paid, and where it falls short.

Read: recovery scores compared, brand by brand

Apple Watch

Readiness · 0–10 · Series 12 and Ultra 4 only · free

Score name: Readiness · Free
How it's built: A 0–10 score launched in September 2026, built from three groups of inputs compared with your own recent baseline: Activity (active calories, workout effort ratings and Training Load), Vitals (overnight heart rate, respiratory rate, wrist temperature, HRV and blood oxygen, plus daytime vitals) and last night's Sleep Score. Labels: Recover 0–1, Pace Yourself 2–4, Ready 5–7, Go for it 8–10. Unlike WHOOP and Oura it updates through the day, after workouts, naps and changes in daytime heart rate. It needs at least 7 nights of sleeping with the watch, against a baseline of 7 of the past 49 nights. Series 12 and Ultra 4 only: older watches on watchOS 27 do not get it and still rely on third-party apps such as Athlytic, Bevel, Gentler Streak and Training Today.
Scale: 0–10 · Recover 0–1 / Pace Yourself 2–4 / Ready 5–7 / Go for it 8–10
Core inputs: Activity (active calories, workout effort, Training Load), overnight and daytime vitals (heart rate, HRV, respiratory rate, wrist temperature, blood oxygen), Sleep Score
Measurement window: Overnight baseline, updates through the day
Validation: No published validation of the score, from Apple or anyone else. Built on Apple Heart and Movement Study data. Earlier Apple Watch HR and HRV are well validated independently; the Series 12 heart rate study is Apple-sponsored and did not test overnight HRV.

Unique strength

The only mainstream score on a 0–10 scale, it updates through the day, and no subscription is stated. Older watches still have the largest third-party app ecosystem.

Watch out for

Series 12 and Ultra 4 only. No score until 7 nights of sleep data, and a night without the watch can leave you without one. US watches with part numbers ending LW/A or LS/A have no blood oxygen in Vitals, so their score uses one fewer input. No validation published.

How the Apple Watch Readiness score works

View Apple Watch on Amazon

Affiliate link: we may earn a commission.

RingConn

Wellness Balance · 85–100 excellent / 60–84 good · free

Score name: Wellness Balance · Free
How it's built: RingConn markets it as a daily readiness meter, built from four statuses: Sleep, Activity, Relax and Vitals (HR, SpO2, HRV, breathing rate, skin temperature). No weights published. Positioned as an Oura competitor, so users will ask about it.
Scale: 85–100 excellent / 60–84 good / under 60 needs improvement
Core inputs: Sleep, Activity, Relax status, Vitals status (HR, SpO2, HRV, breathing rate, skin temperature)
Measurement window: Updates every 30 minutes
Validation: No validation of the score located. Sleep/HR validated vs PSG (vendor-cited IEEE study).

Unique strength

Sleep and heart rate validated vs PSG (a vendor-cited IEEE study); a low-cost Oura alternative.

Watch out for

A single banded number now, but no weights are published and no validation of the score has been located.

View RingConn on Amazon

Affiliate link: we may earn a commission.

Amazfit / Zepp

HybridCharge · Wake and Evening scores · free app

Score name: HybridCharge · Free
How it's built: In May 2026 Zepp upgraded BioCharge into HybridCharge, and BioCharge is now one of its components. It combines sensor data (sleep, heart rate, resting heart rate, HRV, activity, stress) with Training Load, LifeLoad event logs (stress, illness, alcohol, travel and so on, each with a time and an impact rating) and post-workout RPE. You get a Wake score and an Evening score; logged events count within about an hour, and HybridCharge can prompt you to log context when your signals shift.
Scale: No published scale
Core inputs: Sensor data (sleep, HR, RHR, HRV, activity, stress), Training Load, LifeLoad event logs, post-workout RPE
Measurement window: Continuous + daily log
Validation: Brand-new (May 2026). No validation located for the score or its signals.

Unique strength

The first mainstream score to formally blend sensor data with structured subjective self-report.

Watch out for

Brand-new (May 2026) with no validation.

View Amazfit / Zepp on Amazon

Affiliate link: we may earn a commission.

Oura Ring

Readiness · 0–100 · 9 contributors · score free, details need membership

Score name: Readiness · Paid
How it's built: A 0–100 morning score built from NINE contributors: resting heart rate, HRV balance (past 14 days vs your 3-month average), body temperature, recovery index (how much sleep you got AFTER your HR hit its overnight low; 6h+ is optimal), sleep, sleep balance, sleep regularity, previous-day activity, and activity balance. "Long-term" means your past two months. Ratings: 85+ optimal, 70–84 good, 60–69 fair, under 60 pay attention.
Scale: 0–100 · 85+ optimal / 70–84 good / 60–69 fair / 0–59 pay attention
Core inputs: RHR, HRV balance, body temp, recovery index, sleep, sleep balance, sleep regularity, prev-day activity, activity balance
Measurement window: Overnight
Validation: Readiness correlated with subjective recovery in endurance athletes (2025). Signals most accurate tested: Dial 2025, HRV CCC 0.97–0.99.

Unique strength

Finger-site temperature plus nine contributors is the richest input set in the market, and Oura's signals are the most accurate tested (Dial CCC 0.97–0.99). The only score with even moderate independent support.

Watch out for

The score is visible free; contributors, trends and insights need a membership ($5.99/mo or $69.99/yr). Validation is "moderate," and some of the supporting work is Oura-affiliated.

View Oura Ring on Amazon

Affiliate link: we may earn a commission.

WHOOP

Recovery · percentage (green/yellow/red) · membership

Score name: Recovery · Paid
How it's built: A single morning percentage built from HRV (weighted above sleep), resting heart rate, sleep performance, and respiratory rate, plus blood oxygen and skin temperature on WHOOP 4.0 and newer, all measured during sleep and scored against your own personal baseline rather than population norms. Color bands: green 67–99%, yellow 34–66%, red 1–33% (the average member sits around 58%). Designed to be paired with Strain: "match today's strain to your recovery."
Scale: Percentage · green 67–99 / yellow 34–66 / red 1–33
Core inputs: HRV, resting heart rate, sleep performance, respiratory rate; blood oxygen and skin temperature also included (WHOOP 4.0 and newer)
Measurement window: Calculated once each morning from sleep data
Validation: Tested in NCAA swimmers (Lundstrom 2024): raw HRV carried the signal, the composite did not. Signals OK (Dial 2025, HRV CCC 0.94).

Unique strength

A lean input model, transparent by consumer standards, and Recovery is one of only two scores ever tested against an external reference.

Watch out for

Requires a paid membership. When tested in NCAA swimmers, the raw HRV carried the signal; the composite Recovery score did not add to it.

View WHOOP on Amazon

Affiliate link: we may earn a commission.

Polar

Nightly Recharge · ANS Charge −10 to +10 · free

Score name: Nightly Recharge · Free
How it's built: Nightly Recharge combines ANS Charge and Sleep Charge. ANS Charge is the autonomic piece: heart rate, HRV, and breathing measured over the first ~4 hours of sleep versus your 28-day usual level. Three nights to baseline. Polar Loop supported.
Scale: ANS Charge −10 to +10 (plus a Sleep Charge component)
Core inputs: HR, HRV, breathing over the first ~4h of sleep, vs a 28-day baseline
Measurement window: First 4h of sleep
Validation: Sleep staging validated vs PSG, but not the composite. Nocturnal RHR poor (Dial 2025, CCC 0.86).

Unique strength

The 28-day overnight baseline is one of the most personalized windows in the market; pairs with the H10 chest strap (the lab reference for most wearable studies).

Watch out for

Wrist nocturnal RHR/HRV were poor in Dial 2025 (RHR CCC 0.86). No independent validation of the composite.

View Polar on Amazon

Affiliate link: we may earn a commission.

Samsung

Energy Score · 0–100 · Galaxy Watch 4 or newer and Galaxy Ring · free

Score name: Energy Score · Free
How it's built: A 0–100 overnight score from activity, sleep, sleeping heart rate, and sleeping HRV, with age and gender weighting the factors. Activity uses an Acute:Chronic Workload Ratio; sleep uses an "energy reservoir" model. Built with Prof. Patrick O'Connor (University of Georgia) as "Overall Capacity": physical plus cognitive. Needs 24h of wear for a first score; 7 days recommended. Galaxy Watch 4 or newer and Galaxy Ring (Samsung Health app).
Scale: 0–100
Core inputs: Activity, sleep, sleeping HR, sleeping HRV (age & gender weighted)
Measurement window: Overnight + 24h wear
Validation: Only company-affiliated work (UGA, cognitive performance). No independent score test.

Unique strength

Built with an academic collaborator and free across the Galaxy Watch line.

Watch out for

The only validation is company-affiliated (UGA correlated it to cognitive-test performance); no independent test of the score.

View Samsung on Amazon

Affiliate link: we may earn a commission.

Ultrahuman

Dynamic Recovery · 0–100 · re-adjusts intraday · free

Score name: Dynamic Recovery · Free
How it's built: A 0–100 score from five factors (Sleep Index, Stress Rhythm, Temperature, RHR, and HRV Form) that re-adjusts through the day rather than locking a single morning number. Stress Rhythm can move it up to about 10 points either way during the day. Above 85 is optimal, 67 to 85 high. No mandatory subscription, which is its wedge against Oura.
Scale: 0–100 · above 85 optimal / 67–85 high
Core inputs: Sleep Index, resting heart rate, HRV Form, skin temperature, Stress Rhythm
Measurement window: Overnight + daytime
Validation: No independent score validation located. Some signal-level HRV validation only.

Unique strength

Re-adjusts intraday and has no mandatory subscription.

Watch out for

No independent validation of the score located; only signal-level HRV validation.

Fitbit / Pixel

Readiness · 1–100 · free

Score name: Readiness score (Google Health app) · Free
How it's built: A 1–100 overnight score, now in the Google Health app, from three inputs: HRV (from sleep), sleep over the past week, and resting heart rate. The old "activity" input was removed and replaced by RHR. Seven nights to set a baseline (about a month for the full picture). Labels: low 1–29, moderate 30–64, high 65–100. Google changed the calculation in the Google Health app, so past scores may shift. Free in the base Google Health app; Premium adds coaching built on the score.
Scale: 1–100 · low 1–29 / moderate 30–64 / high 65–100
Core inputs: HRV (from sleep), sleep (past week), resting heart rate
Measurement window: Overnight
Validation: No published independent test of the score. Signal-level validation only.

Unique strength

Free, simple, three-input model.

Watch out for

No published independent test of the score ("built on research"). Signal validation is moderate and component-level only.

View Fitbit / Pixel on Amazon

Affiliate link: we may earn a commission.

Garmin

Training Readiness · 1–100 · Body Battery is separate · free

Score name: Training Readiness · Free
How it's built: A 1–100 score, updated continuously, that blends your overnight recovery with your training workload: sleep score, recovery time, HRV status, acute load, sleep history (3 nights), and stress history (3 days). Body Battery is a SEPARATE Firstbeat energy gauge, a common point of confusion. The 6th Training Readiness factor is sleep history, NOT Body Battery. Available on many Forerunner, fēnix, Epix, Venu and Descent models.
Scale: 1–100 · Prime 95–100 / High 75–94 / Moderate 50–74 / Low 25–49 / Poor 1–24
Core inputs: Sleep score, recovery time, HRV status, acute load, sleep history (3 nights), stress history (3 days)
Measurement window: Continuous
Validation: Firstbeat white papers cover components, not the composite. Nocturnal HRV poor (Dial 2025).

Unique strength

The most training-load-oriented of the mainstream scores. Built for athletes managing workload, and free.

Watch out for

No independent validation of the composite. Garmin's nocturnal HRV/RHR were poor in Dial 2025 (excluded from the RHR analysis on methodology grounds).

View Garmin on Amazon

Affiliate link: we may earn a commission.

Suunto

Recovery · % · HRV, sleep, training balance · free

Score name: Recovery · Free
How it's built: A percentage recovery score built from your 7-day average HRV, 7-day sleep and Training Stress Balance. On Suunto Run, plus Race, Race S and Vertical (Aug 2025) and 9 Peak Pro (Sep 2025). Body Resources is a separate heart-rate-based gauge.
Scale: Percentage
Core inputs: 7-day average HRV, 7-day sleep, Training Stress Balance
Measurement window: Daily
Validation: No independent validation located.

Unique strength

Blends a week of HRV and sleep with training balance, for outdoor and endurance athletes.

Watch out for

No independent validation located.

View Suunto on Amazon

Affiliate link: we may earn a commission.

COROS

Recovery % · training load only · free

Score name: Recovery · Free
How it's built: The genuine outlier: no physiological recovery score. Recovery is a percentage built from base fitness, training load and the energy remaining since your last workout, updated hourly. Purely training-load and fitness math: no HRV, temperature, or sleep in the model. COROS measures overnight HRV and a Sleep Score (Aug 2026) separately; neither feeds Recovery.
Scale: 0–19% exhausted / 20–69% fatigued / 70–89% normal / 90–100% fresh
Core inputs: Base fitness, training load, remaining energy since last workout. HRV and sleep are tracked but excluded
Measurement window: Real-time (hourly)
Validation: Not a recovery score, so nothing to validate as one. No HRV input.

Unique strength

Honest, performance-framed training-load math for endurance athletes, no "wellness" overreach.

Watch out for

Not a physiological recovery score at all, so there is nothing to validate as one. The De Gruyter paper lists its composite as "Daily Stress."

View COROS on Amazon

Affiliate link: we may earn a commission.

Can you trust the number?

2 of 12 scores have any independent validation.

The scores are mostly unvalidated black boxes; the underlying HRV/RHR signals are validated in some devices. Only WHOOP and Oura have been tested against a reference, and WHOOP's came out weak. Trust the signal, not the score.

Read: can you trust your recovery score?

Score vs signal

Is it validated: the score, or just the signals?

"None located" is a real finding, not a gap: each brand was searched individually for a study testing its score.

12 brands

Brand & score	Score validated?	Signals (HRV/RHR)?	Key evidence
WHOOP Recovery	Tested: weak	Moderate	Tested in NCAA swimmers (Lundstrom 2024): raw HRV carried the signal, the composite did not. Signals OK (Dial 2025, HRV CCC 0.94).
Oura Ring Readiness	Moderate	Strong	Readiness correlated with subjective recovery in endurance athletes (2025). Signals most accurate tested: Dial 2025, HRV CCC 0.97–0.99.
Garmin Training Readiness	None located	Poor	Firstbeat white papers cover components, not the composite. Nocturnal HRV poor (Dial 2025).
Fitbit / Pixel Readiness score (Google Health app)	None located	Moderate	No published independent test of the score. Signal-level validation only.
Samsung Energy Score	None located	Validated	Only company-affiliated work (UGA, cognitive performance). No independent score test.
Polar Nightly Recharge	None located	Poor	Sleep staging validated vs PSG, but not the composite. Nocturnal RHR poor (Dial 2025, CCC 0.86).
Ultrahuman Dynamic Recovery	None located	Moderate	No independent score validation located. Some signal-level HRV validation only.
COROS Recovery	n/a	n/a	Not a recovery score, so nothing to validate as one. No HRV input.
Amazfit / Zepp HybridCharge	None located	None located	Brand-new (May 2026). No validation located for the score or its signals.
Suunto Recovery	None located	n/a	No independent validation located.
Apple Watch Readiness	None located	Validated	No published validation of the score. Signals well validated at the signal level; the Series 12 heart rate study is Apple-sponsored and not peer reviewed.
RingConn Wellness Balance	None located	Validated	No validation of the score located. Sleep/HR validated vs PSG (vendor-cited IEEE study).

The keystone paper

Doherty, Baldwin, Altini et al., "Readiness, recovery, and strain: an evaluation of composite health scores in consumer wearables" (Translational Exercise Biomedicine, De Gruyter, 2025) catalogued 14 composite scores across 10 manufacturers. Co-author Marco Altini is an HRV scientist and an Oura advisor, so this is not an anti-wearable piece. Verbatim: "None of the manufacturers disclosed their exact algorithmic formulas, and few provided empirical validation or peer-reviewed evidence supporting the accuracy or clinical relevance of their scores." Most common inputs: HRV 86%, RHR 79%, activity 71%, sleep 71%.

Eight headlines from the data

If you only remember eight things.

What you're buying

One number

A modeled opinion, not a measurement

Almost every brand now sells a single morning number for "how recovered are you." But there's no clinical gold standard for recovery the way polysomnography exists for sleep, so brands validate their inputs, not the output. A recovery score is a modeled opinion layered on real signals.

Cross-brand comparison

Don't

An "80" means different things

Different inputs, baselines, and scales. Polar isn't even on 0–100 (it's ANS Charge, −10 to +10). Two devices on one wrist will routinely disagree. Only the trend versus your own baseline is meaningful, never the cross-brand number.

Scores independently validated

2 / 12

Only WHOOP & Oura

And WHOOP's came out weak: the raw HRV carried the signal, the composite didn't add to it. Everyone else has at most component validation or vendor white papers. "2 of 12" is the headline finding, not a coverage gap.

The common spine

HRV+RHR+sleep

Same ingredients, different recipe

Across the De Gruyter catalogue: HRV in 86% of scores, RHR 79%, activity 71%, sleep 71%. The differences between brands are weighting and window length, not the raw ingredients.

Biggest controllable hit

~15%

Alcohol drops HRV overnight

Oura's 600,000-member data: HRV down ~15% (~10.8 ms), HR up ~9.6%. WHOOP: a single drink cuts next-day recovery ~8%. Dose-dependent, and the clearest lever a food-logging user can actually see and change.

Single biggest lever up

Sleep

Consistency = duration

Adequate, regular sleep is the dominant positive input across every brand. Regularity matters as much as hours. Everything else (supplements, cold plunges, breathwork) is a rounding error next to it.

Signals beat scores

Oura best

Cite the signal, not the score

Dial 2025 (536 nights vs ECG): Oura HRV CCC 0.97–0.99, WHOOP 0.94, Garmin and Polar poor. The ingredients are reasonably sound (best in Oura); the recipe layered on top is the unvalidated part.

The independence problem

Double-count

One bad night hits 3 inputs

Poor sleep lowers next-day HRV and raises RHR, so a score "combining" sleep + HRV + RHR partly counts the same event twice. This can overstate a single bad night, a known critique of every composite score.

Want every brand on every metric in one place? Open the wearable accuracy hub

Keep exploring

Related tools.

More free, evidence-based tools to get the most out of your wearable.

HRV, beat-to-beat

HRV Factor Explorer

43 research-backed factors that affect Heart Rate Variability, ranked by evidence.

Recovery · 43 factors

Open

VO2 max estimate

VO2 Max Factor Explorer

39 research-backed factors that raise or lower VO2 max, ranked by evidence strength.

Recovery · 40 sources

Open

Fitbit Air vs WHOOP

Heart-rate, sleep and calorie accuracy with numbers attached, plus the 3-year cost.

Wearables · head-to-head

Open

Common questions

Recovery scores, answered.

Can I compare a recovery score across different brands?

No. Every brand uses different inputs, baselines, and scales, so a "75" on one device does not mean a "75" on another. Polar is not even on a 0–100 scale (ANS Charge runs −10 to +10). Two devices on the same wrist will routinely disagree. Only the trend versus your own personal baseline is meaningful, never the cross-brand number.

Are wearable recovery scores actually validated?

Mostly no. Of 12 brands, only WHOOP and Oura have ever been tested against an external reference, and WHOOP's came out weak: the raw HRV carried the signal while the composite Recovery score added no consistent association. Oura has moderate support. Everyone else has at most component validation or vendor white papers, not an independent test of the score itself. There is no clinical gold standard for "recovery" the way polysomnography is for sleep, so brands validate their inputs, not the output. A recovery score is a modeled opinion, not a measurement.

Which brands have a free recovery score?

Free: Garmin Training Readiness, Fitbit/Pixel Readiness, Samsung Energy Score, Polar Nightly Recharge, Ultrahuman Dynamic Recovery, COROS Recovery, Amazfit HybridCharge, Suunto Recovery, and RingConn. Paid: WHOOP (band requires a membership). Oura's score is visible free, but its breakdown and trends need a membership ($5.99/mo). Apple Watch Readiness is free with no subscription stated, but only on Series 12 and Ultra 4; older Apple Watches still need a third-party app.

What lowers your recovery score the most?

Alcohol is the biggest single controllable hit: Oura's 600,000-member data shows HRV down about 15% (~10.8 ms) and heart rate up about 9.6%, and WHOOP found a single drink cuts next-day recovery roughly 8%. Sleep loss, hard training or overreaching, and the onset of illness are the other strong-evidence drops. Late or high-glycemic meals near bed, evening caffeine, dehydration, altitude, and nicotine are moderate hits.

Does the Apple Watch have a recovery score?

Yes, since September 2026, but only on Apple Watch Series 12 and Ultra 4. Apple's Readiness score runs from 0 to 10 (Recover, Pace Yourself, Ready, Go for it) and combines your recent activity and Training Load, your overnight and daytime vitals, and your Sleep Score, all against your own baseline. It updates through the day and needs at least 7 nights of sleep data before it appears. Apple has published no validation of the score. Older Apple Watches still rely on third-party apps like Athlytic, Bevel, Gentler Streak and Training Today.

Which supplements actually help recovery?

The defensible ones, by evidence: magnesium (bisglycinate), glycine, tart cherry, L-theanine, ashwagandha, and saffron. Omega-3 (DHA) modestly lowers resting heart rate. Melatonin is a circadian timing aid, not a recovery booster. Do not rely on the trendy isolated apigenin (it has no human sleep RCT; the evidence is for whole chamomile extract), or on valerian, GABA, CBD, or creatine as recovery aids. Trial quality across this whole category is often low; doses are study doses, not recommendations, and this is not medical advice.

Why do my two wearables show different recovery numbers?

Because they are not measuring the same thing the same way. The common spine is HRV + resting heart rate + sleep differenced against a personal baseline, but each brand weights those differently, uses a different window (Polar reads only the first ~4 hours of sleep; Garmin and Coros lean on training load), and anchors to a different baseline length. The scores are inferences, not measurements, so disagreement is expected.

How does what I eat and drink affect my recovery score?

Alcohol, late or heavy meals within ~3 hours of bed, high-glycemic spikes near bedtime, evening caffeine, and dehydration all move the HRV + RHR + sleep signals your score is built on. These are the levers a food-logging user can actually see and change, and the brands' own blogs tend to underplay them, which is exactly why connecting nutrition to your recovery trend is the highest-value angle.

Sources

Every brand checked against vendor docs; every factor and supplement claim verified against primary research.

Recovery & composite-score validation

De Gruyter 2025, Readiness, recovery & strain: evaluation of composite health scores

Doherty / Altini et al.

Recovery & composite-score validation

Marco Altini: summary of the composite-scores paper

Substack

Recovery & composite-score validation

Dial et al. 2025: nocturnal RHR/HRV validation, 536 nights vs ECG

Physiological Reports

Recovery & composite-score validation

Bellenger et al. 2021: WHOOP wrist-PPG HR/HRV validation vs ECG

Sensors, MDPI

Recovery & composite-score validation

Lundstrom et al. 2024: WHOOP metrics vs energy deficiency in elite swimmers

Sage

Recovery & composite-score validation

Penn State thesis: WHOOP score vs metabolism/recovery in D1 swimmers

etda.libraries.psu.edu

Recovery & composite-score validation

CTS / TrainRight: analysis of the swimmer wearable-score study

trainright.com

Recovery & composite-score validation

Sports Medicine Open 2025: validating subjective ratings with Oura data in elite endurance athletes

link.springer.com

Recovery & composite-score validation

SciTePress 2025: Recovery & Readiness Monitoring Using Wearable Technology

scitepress.org

Brand documentation

WHOOP, Recovery: How It Works

Recovery 101

Brand documentation

Oura: Readiness Contributors

Oura Help

Brand documentation

Garmin: Training Readiness

owner's manual

Brand documentation

Fitbit / Google: Readiness score

Google Health Help

Brand documentation

Samsung × University of Georgia: Energy Score

Samsung Newsroom

Brand documentation

Polar: Nightly Recharge / ANS Charge

Polar Support

Brand documentation

Ultrahuman: Recovery score explainer

Ultrahuman Science

Brand documentation

COROS: EvoLab Metrics

COROS Help

Brand documentation

COROS: Understanding the Recovery Widget

COROS Help

Brand documentation

Zepp: What is HybridCharge?

Zepp OS news

Brand documentation

Suunto: Recovery widget

Suunto Run support

Brand documentation

Apple: Track Your Training Load on Apple Watch

Apple Support

Brand documentation

Apple: Understand your readiness score on Apple Watch

Apple Support

Brand documentation

Apple: Use the Readiness app on Apple Watch

watchOS User Guide

Brand documentation

RingConn: Health Scores Explained

RingConn