In This Article
The short answer: When your HRV looks fine but your sleep score is poor, or when HRV is suppressed but you slept well, the recovery score typically follows the worse of the two signals. This is the right call: your body recovers as a system, and a weak link in any input limits total recovery capacity. Rather than averaging signals, look at each metric individually to understand what specifically went wrong that night.
Read key takeaways →
How Recovery Scores Are Built
A recovery score (Oura calls it Readiness; WHOOP calls it Recovery) is an algorithm-generated composite number. It aggregates several biological signals measured overnight into a single percentage or score. The specific inputs vary by device, but the major components across most platforms are: HRV (the night before and 7-day rolling baseline), resting heart rate, sleep duration, sleep stage distribution (particularly deep sleep and REM), body temperature deviation, and sleep timing relative to your habitual schedule.
None of these inputs are directly measured in the same biological system. HRV reflects autonomic nervous system state. Resting heart rate reflects cardiac recovery. Sleep stages reflect neural recovery processes. Skin temperature reflects immune and thermoregulatory state. A conflict between these metrics is not a flaw in the system: it is a real signal that different recovery systems are in different states. The question is which system is currently the limiting factor.
What Each Input Reflects
The Common Conflict Patterns
Most metric conflicts follow recognizable patterns. Each has a specific physiological explanation and a specific recommended response.
Good HRV, poor sleep score
Your autonomic system recovered well (sufficient deep sleep quality and low psychological stress), but total sleep duration, timing, or sleep efficiency was poor. You may have gone to bed late, woken early, or had fragmented sleep that reduced deep sleep and REM percentage. Readiness follows the sleep quality signal here, rightly so. Adenosine clearance and growth hormone release were incomplete. Even if your HRV is green, chronic sleep fragmentation compounds over time.
Poor HRV, good sleep score
You slept well structurally (good duration, good staging), but your HRV is suppressed. The most common cause: systemic stress that predates sleep. Heavy training the previous day, alcohol from the evening before, accumulated psychological stress, or early-stage immune activation. Sleep cannot fully restore HRV within one night when the underlying stressor is still active. The recovery score correctly identifies this as a suboptimal day.
Good HRV and sleep, elevated skin temperature
Your autonomic and sleep signals look fine, but skin temperature is running above baseline. The most likely cause: alcohol from the previous evening (its effects clear within 24 hours), the luteal phase of the menstrual cycle, or a warm sleep environment. If skin temperature is elevated alongside normal HRV and good sleep, this is often a false alarm on the readiness score rather than a genuine recovery problem. Understand the cause before reducing training.
Everything suppressed simultaneously
When HRV is down, sleep quality is poor, resting heart rate is elevated, and skin temperature is above baseline, this is a genuine, multi-system recovery deficit. The most common causes: emerging illness, accumulated training overload (functional overreaching), or a combination of alcohol, poor sleep, and high stress. Do not train hard. Prioritize sleep and nutrition. Track the pattern over the following 2-3 days.
The Decision Framework: What to Do With Conflicting Data
The fundamental principle: when metrics conflict, look for the weakest signal and let it inform your training decision. Your body recovers as a system. A strong HRV score does not compensate for a sleep deficit that left significant adenosine uncleared. A good sleep score does not override an autonomic system still processing yesterday's stress load.
Decision Cards by Pattern
HRV green + sleep good + temperature normal
Train as planned. All major recovery systems are ready. This is the day for demanding work.
HRV green + sleep poor (fragmented or short)
Moderate the session. Your autonomic system is ready but neural recovery (adenosine clearance, growth hormone) was incomplete. Avoid max-intensity work. A solid moderate session is fine.
HRV suppressed + sleep good + temperature normal
Reduce intensity by 20-30%. Your autonomic system is still processing yesterday's load. Low to moderate training is fine and will not impair recovery. Avoid intervals or heavy lifts.
HRV green + sleep good + temperature elevated (known cause)
Proceed with normal training if the temperature cause is known and benign (alcohol cleared, luteal phase, warm room). If the cause is unknown, treat as a yellow day.
HRV suppressed + sleep poor + temperature elevated
Rest or very light movement only. Multi-system deficit, possible illness onset. Pushing through a session in this state extends recovery time rather than shortening it.
The most important skill is identifying which signal is the reliable one and which is the confounded one. Skin temperature is most often confounded by environment and alcohol. HRV is most often confounded by residual psychological stress. Sleep score is most often confounded by a single late night rather than a chronic pattern. Interpreting conflicts requires knowing the context around each night.
Common Misconception
A single night of conflicting data does not define your recovery state. One outlier night, whether caused by alcohol, a late work session, an anxiety spike, or a warm room, can produce a misleading recovery score. The pattern across 3-5 days is far more informative than any single morning's readout. Do not make significant training program changes based on one bad score.
Why HRV Is the Most Reliable Anchor
When you are uncertain which metric to trust in a conflict, default to the 7-day HRV trend rather than any single-night reading. Your 7-day HRV baseline is the most robust signal available from consumer wearables. It averages out single-night noise from environmental factors, position changes, and measurement artifacts, and reflects the actual state of your autonomic nervous system as a rolling average.
Plews et al. (2013, International Journal of Sports Physiology and Performance) demonstrated that the coefficient of variation (CV) of HRV measured across 7 days is substantially lower than single-night readings, meaning the 7-day average is a more stable and accurate reflection of recovery status. This is why devices like Oura report your HRV against your 7-day rolling baseline rather than against population norms.
HRV Reading Rules
- →Single-night HRV: High variability. Use as context only. Do not make training decisions on a single-night HRV number alone.
- →7-day HRV trend: The reliable signal. A declining 7-day trend indicates accumulating stress. A stable or rising trend indicates recovery is keeping pace with load.
- →Deviation from baseline: More useful than absolute number. Your HRV at 90% of your 7-day baseline is a clear signal regardless of whether your absolute HRV is 40ms or 80ms.
For the full framework on reading HRV data, including how to use the baseline comparison and when suppressed HRV requires action, see the HRV interpretation guide. For how HRV fits into the full recovery decision framework, see the Recovery Protocol.
Frequently Asked Questions
Should I trust the overall score or look at individual metrics?
Both, in sequence. The overall score is your first filter: green means probably fine, red means investigate. Then look at each underlying metric to understand what drove the score. A red score from temperature elevation alone (alcohol the night before) requires different action than a red score from both suppressed HRV and poor sleep staging. The aggregate score tells you something happened; the individual metrics tell you what.
What if my HRV is chronically low even when I sleep well?
Chronic HRV suppression despite good sleep usually indicates an ongoing systemic stressor that sleep alone cannot resolve: chronic psychological stress, high training load without adequate recovery days, persistent low-grade inflammation, or inadequate calorie or carbohydrate intake for your training volume. HRV responds to the total stress load on the body, not just training stress. If HRV has been trending down for 2+ weeks and sleep quality is not the issue, look at training volume, nutritional intake, and psychological stress as the more likely culprits.
How do I know if a score conflict is a device error or a real signal?
Device errors typically produce clearly implausible readings: an HRV of 200ms, a resting heart rate of 30 bpm, or sleep data showing 0% deep sleep for multiple nights. These are measurement artifacts from poor device fit, movement, or sensor malfunction. Metric conflicts within plausible biological ranges (HRV in your normal window but sleep score below average) are almost always real physiological signals, not device errors. If you see implausible readings, check device fit and charging status first.
Is a readiness score useful if I train at the same time every day regardless?
Yes, for a different reason. Even if you train on a fixed schedule, your readiness score tells you how to calibrate the session intensity. A green readiness day is the right day to push hard, test a new weight, or run a time trial. A red day is the right day to do the session at 70% effort and focus on movement quality rather than output. Using readiness to modulate intensity within a fixed training schedule is one of the highest-leverage applications of the data.
What does it mean when my sleep score is great but I feel exhausted?
This is usually a case where the device is measuring sleep duration and staging accurately, but missing something the sensor cannot capture: psychological or emotional exhaustion, accumulated life stress, or a recovery debt that precedes the night being measured. Wearables measure physiological signals during sleep; they do not measure the subjective experience of recovery or the psychological weight your nervous system is carrying. If subjective fatigue persistently diverges from your sleep score, your stress load outside of sleep is the likely explanation. Also check HRV trend: if it is declining despite good sleep scores, the cumulative stress load is exceeding what sleep alone can clear.
What to Remember
- →A recovery score is a composite of multiple independent biological signals. Conflicts between metrics are not errors; they reflect real differences in recovery state across different physiological systems.
- →When metrics conflict, let the weakest signal guide your training decision. Your body recovers as a system. Strong HRV does not compensate for incomplete sleep-based adenosine clearance.
- →HRV trend over 7 days is the most reliable anchor metric. Single-night HRV readings have high variability. Base training decisions on the trend, not any one morning number.
- →Skin temperature is the most commonly confounded metric. Known causes (alcohol, luteal phase, warm room) should be ruled out before treating temperature elevation as a recovery signal.
- →Multi-system suppression (HRV down, sleep poor, temperature elevated, resting HR high) warrants genuine rest. This pattern often precedes illness or training overreach and should not be pushed through.
- →A single conflicting night is noise. The pattern across 3-5 days is signal. Do not make major training adjustments based on one outlier morning readout.
Related on Protocol
How to Interpret Your HRV Data
The baseline comparison framework, what causes HRV to drop, and how to use it as a primary training signal.
What Your Skin Temperature Data Actually Tells You
How skin temperature deviation works as an early warning signal for illness and why it is the most commonly confounded wearable metric.
Why Your Recovery Score Changes Day to Day
What goes into the readiness score algorithm and what each component reflects about your recovery state.
Protocol
See the full recovery picture, not just one number
Protocol surfaces HRV, sleep quality, skin temperature, and resting heart rate together so you can identify what is driving your recovery score and what to do about it.
Get started freeNo credit card required.
References
Key Researchers
- Daniel Plews — Auckland University of Technology Sports scientist specializing in HRV monitoring methodology. Research on HRV measurement reliability, 7-day rolling baselines vs. single-night readings, and practical implementation for athletes.
- Bruce McEwen — Rockefeller University Neuroendocrinologist. Foundational work on allostatic load, the cumulative physiological cost of stress across multiple systems, which explains why recovery metrics from different biological systems can diverge.
- Matthew Walker — University of California, Berkeley Sleep scientist. Research on sleep stage function, adenosine clearance, and the consequences of incomplete sleep for cognitive and physical recovery, underpinning why sleep score conflicts with HRV are physiologically meaningful.
Key Studies
- Plews et al. (2013) — HRV measurement reliability International Journal of Sports Physiology and Performance. Demonstrated that the 7-day rolling HRV average is significantly more reliable than single-morning readings, supporting trend-based interpretation over single-night data.
- Buchheit (2014) — Monitoring training status with HRV Sports Medicine. Comprehensive review of HRV as a training monitoring tool, including how to interpret HRV alongside other recovery inputs and when conflicts between metrics indicate genuine physiological divergence.
