EVIDENCE-BASED
Built on peer-reviewed
sports science.
Every threshold in COMPASS (the 0.5°C clearance, the R score weighting, the 6-scan baseline transition) is anchored to published literature. No guesswork, no proprietary black boxes.
THE TWO AXES
These questions do not overlap.
Systemic Readiness
Scale: 0–100
“Is this athlete's body ready to perform today?”
Derived from wearable biometrics via the Spike aggregation layer. Reflects systemic capacity: sleep quality, HRV recovery, resting heart rate, and acute-to-chronic workload ratio.
Thermal Asymmetry Index
Unit: °C bilateral delta
“Is there developing tissue stress in a specific body region?”
Derived from NUR Scanner thermal imaging + NURAI computer vision. Identifies contralateral asymmetry: the temperature difference between left and right sides of each region.
“A wearable cannot detect that the right hamstring is inflamed. A thermogram cannot tell you whether the athlete slept well. The clinical value lives in their non-overlap, not their overlap.”
COMPASS Founding Principle
COMPASS SCORING
Every threshold is deliberate.
Zone floors are anchored to published asymmetry intervention thresholds and validated systemic readiness cut-points.
76–100
R ≥ 50 · T < 50
Full training load approved. No capacity gap.
Scan: 7–10 days
51–75
R < 50 · T < 50
Volume-reduced training, sleep and HRV optimisation.
Scan: 5–7 days
26–50
R ≥ 50 · T ≥ 50
Mandatory physiotherapy assessment before load increase.
Scan: 3–4 days
0–25
R < 50 · T ≥ 50
Mandatory clinical intervention. Remove from training immediately.
Scan: Within 3 days
PEER-REVIEWED RESEARCH
Six papers. One framework.
The COMPASS scoring engine is grounded in published sports science, not proprietary algorithms.
Gómez-Carmona et al.
“83% of soft tissue injuries identified 8.2 days before first symptom”
Foundational justification for T score threshold and early-warning capability
Arnaiz-Lastras et al.
“64% reduction in muscle injuries with thermal prevention protocol at ≥0.4°C asymmetry”
Establishes the 0.4°C–0.5°C intervention threshold used in COMPASS scoring
Schiavon et al.
“Meta-analysis: AUC 0.83, sensitivity 0.81, specificity 0.74”
Validates thermal imaging as a reliable injury detection modality at population scale
Claudino et al.
“Individual baseline ML models: AUC 0.84 vs 0.61 for group-norm baselines”
Justifies the 6-scan transition from population to personal 1.5 SD baseline
Kalkhoven, Watsford & Impellizzeri
“Systemic load measures cannot quantify tissue-specific injury risk”
Foundational theory: establishes why thermal and wearable axes are non-overlapping
ThermoHuman
“Population mean 0.004 ± 0.66°C across 950 healthy athletes”
Calibrates the population-level baseline used during the first 6 scans
INDIVIDUAL BASELINE
The system learns each athlete.
Asymmetry thresholds should not be static across all athletes. NUR transitions from population-level norms to a personal 1.5 SD baseline after 6 qualifying scans.
Calibrated against ThermoHuman database: 950 healthy athletes, mean 0.004 ± 0.66°C. Initial zone assignments use this normalisation.
1.5 SD of the athlete's own scan history (~0.4°C for most athletes). Personal fluctuation patterns are removed from the signal, increasing specificity.
RTP CLEARANCE PROTOCOL
Two consecutive scans. No exceptions.
Return to Play clearance requires the affected region to show bilateral asymmetry ≤ 0.5°C in two consecutive qualifying scans, replacing time-based protocols with objective data.
Eligible for clearance: 2 consecutive scans below 0.5°C
Threshold derived from: Gómez-Carmona et al., 2020; Arnaiz-Lastras et al., 2019
NUR PLATFORM