The science behind every number Tuwa shows you
No black boxes. Here is exactly how the readiness score and training-load model work, where the data comes from, and what they can and can't tell you.
How the readiness score is composed
Four physiological signals, weighted against your own baseline
Each morning, Tuwa pulls overnight physiological data from Apple HealthKit and combines it with a short wellness check-in to produce a single readiness number between 0 and 100. The inputs are not treated equally, and none of them is judged against a population norm — every signal is compared to your own recent trend.
Heart-rate variability (HRV — the millisecond-level variation between consecutive heartbeats) is the heaviest-weighted input, because it tracks autonomic nervous-system balance more directly than any other consumer signal. Resting heart rate adds a second, slower-moving view of systemic stress. Sleep duration and consistency capture the recovery window itself. The morning wellness check-in — where you rate soreness, energy, and stress — adds the context no wearable can measure.
Tuwa weights each factor by how far it sits above or below your personal baseline, then blends them. A score of 62 is never just a number: the app tells you that your HRV is tracking 8% below your recent baseline, your sleep ran short, but your resting heart rate is normal. You see the reasoning, not just the verdict.
The training-load model
Acute vs chronic load, ACWR, and the 0.8–1.3 band
Every session you log is converted into a load value from volume and intensity (using RPE — rate of perceived exertion, the 1–10 scale for how hard a set felt — and optionally RIR, reps in reserve). Tuwa then maintains two views of that load: acute load, roughly the last week, which reflects current fatigue, and chronic load, roughly the last four weeks, which reflects your fitness base.
The ratio between them is the acute:chronic workload ratio (ACWR). Research originating with Tim Gabbett and colleagues often uses a relative range around 0.8 to 1.3 as workload context. Below the range, you may be doing much less than your recent base; above it, the current week may be asking for a bigger jump than your recent training supports.
Rather than crude rolling sums, Tuwa computes these loads with exponentially weighted moving averages (EWMA) — a smoothing method that gives recent sessions more weight than older ones. This lets your model respond quickly to a hard block or a deload instead of being anchored to weeks-old data, and it flags spikes before you train, not after.
Data and privacy
On-device by default, offline-capable, composite-only sync
Your raw HealthKit data — every HRV reading, heart-rate sample, and sleep record — stays on your device. Tuwa reads it locally to compute your scores and never uploads the underlying physiological streams to a server.
The app works fully offline. Scoring, load calculation, and spike detection all run on-device, so a flaky connection or a basement gym never interrupts the analysis.
Coach-facing features sync only the composite outputs — your readiness score and load metrics — never the raw biometric data behind them. You share the conclusion, not your nervous system's full record.
Honest limitations
What the model can and cannot do
Tuwa is useful from day one because it starts with sensible population baselines, but its real value is personal. Over roughly the first seven days it learns your own HRV, resting heart rate, and sleep patterns, and the readiness score becomes meaningfully more accurate as those baselines settle. Early scores should be read as directional, not precise.
No model built on consumer wearables is perfect. HRV is noisy — a glass of wine, an early alarm, or a poor night can move a single reading — which is exactly why Tuwa tracks trends rather than reacting to individual data points. The ACWR framework describes population-level workload associations, not individual guarantees or predictions; it informs decisions, it does not make them for you.
Tuwa is a training tool, not a medical device. It does not diagnose illness, injury, or any health condition, and nothing in the app is medical advice. If a metric concerns you or you feel unwell, consult a qualified professional.
Go deeper
Primary sources
Start coaching yourself with confidence
Download Tuwa to plan, log, adapt, and review your strength training in one place.
