The accuracy of every activity recommendation a wearable device produces is constrained by the physiological model underneath it. If the model sees one dimension of a user's fitness, the recommendations will be right for some users and wrong for the rest. The platforms that solve this will build durable user trust. The ones that don't will compete on hardware while their software feels interchangeable.
Every major wearable platform uses a physiological model built around a single threshold measurement. This measurement was not chosen for its physiological depth. It was chosen because it is reproducible in a laboratory. The protocol was designed for researchers, adopted by the industry, and applied identically to every athlete.
The result: two athletes with identical threshold values but fundamentally different physiological profiles receive the same training zones, the same recovery guidance, the same post-workout interpretation. The model cannot distinguish between them. The device treats them as the same person.
The industry has already acknowledged the problem. Over the past decade, every major platform has independently moved away from a single threshold and toward more physiological dimensions. The direction is not in dispute. The question is how far it goes.
Raceproof's Power Threshold Array resolves seven distinct physiological thresholds across the full spectrum of human power output, from neuromuscular through oxidative. Each threshold corresponds to a bioenergetic transition: the point where one energy system yields dominance to another.
Each threshold is individually located using a proprietary model that operates on sensor data wearable devices already capture. PTA requires no new hardware, no new sensors, and no change to the user experience.
PTA is currently built on cycling power data. The same analytical architecture will extend to heart rate, opening the model to every wearable device on the market regardless of sport or sensor type.
| Threshold | Energy system | |
|---|---|---|
| PT-N | Neuromuscular | Peak phosphocreatine sprint capacity |
| PT-S | Sprint | Phosphocreatine depletion onset |
| PT-G | Glycolytic | Anaerobic glycolytic activation |
| PT-X | Maximal | Glycolytic ceiling, maximal anaerobic output |
| PT-O | Oxidative | Oxidative phosphorylation dominance |
| PT-V | VO2max | Maximal aerobic capacity, fully expressed |
| PT-T | Threshold | Functional threshold, lactate equilibrium |
Garmin, Apple, COROS, Polar, Wahoo, and Whoop all perform scalar analysis: a single power value measured at a predetermined duration. Wahoo’s 4DP, the most dimensionally advanced competitor, produces four scalars. But they are still single numbers at durations chosen by the test designer, not located for the individual. The person whose glycolytic peak occurs at 45 seconds and the person whose occurs at 90 seconds both get measured at exactly 60 seconds. Scalar analysis cannot tell them apart.
Raceproof performs vector analysis. Each threshold is resolved as a coordinate pair: the power it produces and the duration at which it occurs. That coordinate is what Raceproof calls the Seat™. Two athletes can produce identical power, but if their thresholds occur at different positions, their physiology is meaningfully different. The Seat captures that difference. No mainstream consumer platform currently packages individualized thresholds as named two-coordinate positions and tracks their movement over time.
The matrix below maps the physiological model depth of every major wearable platform. Each column represents a dimension of the model. PT Seat™, the positional dimension, is exclusive to Raceproof.
The physiological model touches everything the device says to the user. Activity analysis, fitness forecasting, training recommendations, readiness assessment. When the model sees one dimension, every output built on top of it inherits that limitation. The post-workout summary is generic. The suggested workout is generic. The readiness score is generic. The user learns to ignore all of it.
This is not a theoretical problem. The accuracy of readiness and recovery features on major platforms has been publicly questioned by researchers and experienced athletes. When recommendations don't match lived experience often enough, trust erodes. Once trust is lost, the feature is ignored regardless of how prominently it appears on the screen. The underlying model is the limiting factor.
Fourteen physiological signals instead of one or two means the model can distinguish between athletes who look identical under a single-threshold lens but are physiologically different people. Activity analysis becomes specific to each person's energy system engagement. Fitness forecasting reflects development across the full spectrum, not a single number trending up or down. Training recommendations respond to what the individual actually needs, not what the population average suggests. The device stops guessing. It starts knowing.