SSMT Claim Register
SSMT-2026-0004v12026-07-26 Type P · Practice Wearables & Recovery draft
Claim, as marketed
“Your readiness score tells you when to train hard and when to back off.”
Canonical claim
Adjusting training according to a proprietary composite readiness score improves training outcomes compared with not using it.
Scope
Proprietary composite scores — readiness, recovery, body battery, energy score and equivalents. The underlying HR and HRV inputs are a separate matter and are better supported.
Verdict
Unevaluated Never tested, in any design. Currently unevaluable by independent parties.
No independent evidence located as of the card date.
Claim drift — 7 of 7
▲Population▲Setting▲Comparator▲Outcome▲Magnitude▲Reproducibility▲Independence
The first card in the register to drift on all seven axes. Most claims drift on two or three.
Evidence located
14
scores, 10 brands
0
disclose the formula
0
tested acting on it
0
registered trials
Findings

Three layers, and most arguments happen because people stand on different ones.

Layer one — the sensors. Largely fine. Resting heart rate and HRV measured at rest agree closely with ECG (concordance ~0.91 and ~0.94 in independent testing).

Layer two — the score itself. A 2025 review examined every piece of public documentation for 14 composite scores across 10 manufacturers and found that none disclosed its algorithmic formula, with few providing peer-reviewed validation. There is no published criterion standard for 'readiness', so there is nothing to validate the score against. A separate 2025 paper makes the sharper point: proving a device's raw interbeat intervals are accurate does not prove the displayed number is, because undisclosed processing sits in between.

Layer three — acting on it. This is the marketed claim. A 2024 scoping review found eleven studies that adjusted training session-to-session based on readiness; not one used a commercial composite score. A ClinicalTrials.gov search returns no registered trial using a wearable readiness score to prescribe training.

Why reproducibility drifts. Test-retest reliability and cross-device concordance have never been published — nobody has reported whether two devices on the same person on the same night agree on the decision. Scores are also silently versioned: algorithms change between app releases with no changelog, so an athlete's April and July numbers may not be the same measurement.

The inferential error to watch for. The favourable HRV-guided training literature is routinely borrowed to support this claim. That research used supervised chest-strap lnRMSSD with a published decision rule. A black-box 0–100 number is a different intervention.

On the largest dataset. 389 professional golfers over 35,140 monitored nights is a real contribution — but it was authored by the manufacturer's own scientists on the manufacturer's own data, and it is observational. Higher scores co-occurring with better golf is equally consistent with the score being a downstream marker of good sleep and fitness.

What would change this verdict
An independently funded, preregistered RCT allocating participants either to a published, pre-specified decision rule applied to the vendor's score (e.g. <33 deload, >66 proceed) or to an identical periodised plan with the score blinded — hard performance primary outcome, ≥12 weeks, n≥120. Failing that, something far cheaper: put two brands on the same athlete for a month and report how often they disagree. If they disagree often, this grade moves to Contradicted.
Practitioner read
Your readiness score summarises things that genuinely matter — sleep, resting heart rate, HRV — and those inputs are measured reasonably well. But nobody has published how the number is calculated, and no study has ever tested whether changing your training because of it makes you fitter. Use it like the bathroom scale: one more piece of information alongside how you feel and how the warm-up went, not an instruction. If it says back off but you feel great and the session is going well, trust the session.
⚠ Fabricated-source hazard logged. A widely-circulated claim describes a "2024 Frontiers in Physiology study, WHOOP recovery score vs cortisol, n=42, r=0.58". This study does not exist in PubMed; its apparent origin is a content-farm product page. Expect it in AI-assisted right-of-reply submissions.
Could not be verified
Listed rather than dropped. Counts are floors, not censuses.
Sources
  1. Doherty C, et al. (2025). Transl Exerc Biomed 2(2):128–144 — 14 scores, 10 manufacturers ↗
  2. Ibrahim AH, Beaumont CT, Strohacker K (2024). Int J Exerc Sci 17(5):382–404 — scoping review ↗
  3. Dial MB, et al. (2025). Physiol Rep 13(23):e70706 — transparency argument ↗
  4. Grosicki GJ, et al. (2025). Int J Sports Physiol Perform 21(2):180–191 — vendor-authored, n=389 ↗
  5. Miller DJ, et al. (2020). J Sports Sci 38(22):2631–2636 — PSG validation ↗
Every source must resolve at a DOI or PubMed ID.
Changelog
Cite this card
SSMT. "Adjusting training according to a proprietary composite readiness score improves training outcomes compared with not using it." Claim SSMT-2026-0004 v1, graded 2026-07-26. https://sportssciencetech.com/c/SSMT-2026-0004
Disagree with this grade? Every card carries “what would change this verdict” so the specification is explicit. Send evidence and it will be logged and the card re-versioned in public — including when it moves a grade in a manufacturer's favour. Submit evidence →