Research · Recovery

Sleep, HRV, or a Simple Questionnaire? What Actually Predicts Overtraining

By Muhammad Dzulhisham · MSc Sports Coaching, NSCA-CSCS

Every wearable on the market now promises some version of the same claim: a daily score, built from heart rate variability or sleep tracking, that tells you when you're pushing too hard before it costs you a season. It's a compelling pitch. It's also a stronger claim than the actual research supports — according to the people who wrote the field's own consensus statement on the question.

The starting point: no marker has been validated

In 2013, the European College of Sport Science and the American College of Sports Medicine published a joint consensus statement on overtraining syndrome (OTS) — the two largest professional sport-science bodies in the world, co-authored by ten researchers across exercise physiology and sports medicine.¹ Its opening framing is blunt: "Currently, several markers (hormones, performance tests, psychological tests, and biochemical and immune markers) are used, but none of them meet all the criteria to make their use generally accepted." The statement is equally direct later on: "At present, no test meets this criterion" — a specific, sensitive, simple diagnostic test for OTS doesn't exist. The diagnosis, as the authors describe it, is made largely by exclusion: ruling out illness, nutritional deficiency, and other causes of underperformance, not by a positive marker lighting up.

Heart rate variability: promising in theory, inconsistent in practice

HRV is the marker most associated with recovery-tracking wearables today, based on the idea that higher variability reflects more parasympathetic ("rest and digest") tone relative to sympathetic activity. The consensus statement's own assessment: "HRV seems to be a tool in theory but does not provide consistent results. One needs to be careful when using HRV as an outcome measure because there are many different ways to record and calculate the data. Currently, there is no consensus regarding the required standardization and the method of measurement." The studies it reviewed found no change, inconsistent changes, or changes only in specific sub-measures depending on the athlete and protocol.

Simple heart rate numbers hold up slightly better, but with an important caveat about timing:

Most of the short-term signal disappears once training stays elevated for longer than two weeks — exactly the window where knowing you're at risk actually matters.

Hormones: the resting blood draw doesn't work

Cortisol, testosterone, and the testosterone:cortisol ratio have long been proposed as overtraining markers, on the theory that a stressed hypothalamic-pituitary-adrenal axis should show up in resting hormone levels. The consensus statement's conclusion after reviewing the evidence: "resting cortisol is not a useful measurement," and more broadly, "current information regarding the endocrine system and OR/OTS shows that basal (resting) hormone measurements cannot distinguish between athletes who successfully adapt to [overreaching] and those who fail to adapt." Hormonal testing has to be dynamic — repeated exercise-challenge protocols, not a single blood draw — to show anything useful, and even then it's invasive, expensive, and impractical for routine monitoring.

What's actually held up best: asking the athlete, consistently

The tool with the most supportive evidence in the entire consensus statement isn't a wearable metric at all. Subjective mood questionnaires — most studied via the Profile of Mood States (POMS) and the REST-Q Sport inventory — show a genuine, repeatable dose-response relationship with training load: negative mood scores climb as training volume and intensity rise, and recover during tapers, a pattern the literature calls the "iceberg profile." Athletes who go on to develop OTS show a different, more severe pattern than athletes who tolerate the same training load well — most notably a disproportionate rise in depression scores, with some reports finding up to 80% of affected athletes showing signs of clinical depression on these measures. The consensus statement's summary judgment: "research has provided general support for the efficacy of psychological assessments in both basic and applied research involving athletes undergoing overload training."

This isn't a free pass, either — the statement is equally clear that self-report tools carry real risk of "response distortion," athletes consciously or unconsciously answering to present themselves in a better (or occasionally worse) light, especially when they suspect their training load is on the line.

The signal was never missing. The tools marketed as objective — HRV, resting heart rate, cortisol — turn out to be the least consistent ones. The one that holds up is also the simplest: asking the athlete, the same way, regularly, and tracking the trend.

Where sleep actually fits

The consensus statement treats sleep as central to managing fatigue and recovery — but notably not as a diagnostic biomarker to be scored. Its guidance is practical rather than numeric: athletes should sleep for however long it takes to feel wakeful during the day, since individual sleep needs vary considerably and prescribing a fixed number of hours "would be erroneous." The evidence-backed role of sleep here is as an intervention to protect, not a score to optimize — which is a different job than the sleep-stage breakdowns and "readiness" percentages wearables generate, none of which the consensus statement cites as validated OTS markers.

What this means for monitoring, in practice

A real-world example from the consensus statement's own review makes the case for simplicity well: researchers monitoring an entire season of competitive swimmers successfully flagged overtraining using daily training logs — swimmers rated their own fatigue on a 1–7 scale, noted comments about how they felt, and reported illness. Three of the five criteria used to classify a swimmer as overtrained came directly from those subjective daily entries. No wearable, no blood draw.

The practical approach the evidence actually supports is unglamorous: track training load and performance over time, use a simple, consistent subjective wellness questionnaire rather than chasing a single "recovery score," protect sleep as a recovery intervention rather than a metric to game, and — the piece most monitoring tools can't replace — keep an actual conversation open between athlete and coach. That combination is exactly what a testing-first, individualized program is built to do; a device strapped to the wrist wasn't shown to do it better.