Cardiac Drift
Also known as: Heart-Rate Drift, Pa:Hr Decoupling, HR Decoupling
The gradual upward creep in heart rate during a prolonged steady-pace effort, even when speed, power, and perceived effort stay constant. Cardiac drift is the body's progressive struggle to maintain a target intensity as core temperature rises, plasma volume drops, and substrate availability shifts, it's the most reliable physiological signal of accumulating fatigue during a single long session.
Formula
HR drift % = (avg HR second half − avg HR first half) / avg HR first half × 100 [also called "Pa:Hr decoupling" when expressed as pace-to-HR or power-to-HR ratio drift]Example
90-minute Zone 2 run at a steady 5:30/km pace. First-half average HR: 142. Second-half average HR: 152. Drift = (152-142) / 142 = 7.0%. A drift over 5% on a session that's supposed to be aerobic indicates you've slid above Zone 2, the prescribed pace has become too hard for the day's underlying state.
How Afitpilot Uses This
Afitpilot does not currently compute cardiac drift, it requires continuous HR sampling and time-aligned pace/power data, which we don't yet ingest. When endurance-modality TRIMP support lands (`tss-trimp` improvement), drift will be a natural follow-on metric to surface on long aerobic sessions. For now, drift is something endurance athletes track manually via Strava, TrainingPeaks, or Garmin, and is one of the cleanest at-home signals for "is my Zone 2 actually still Zone 2 today?".
Cardiac drift by training state
| Who / Context | Value | Note |
|---|---|---|
| Aerobically well-trained, fresh, cool day | < 3% drift | Cardiovascular system is well-matched to the effort |
| Healthy athlete, average conditions | 3-5% drift | Normal range for a 60-90 min Zone 2 session |
| Mismatched pace (too hard for state) | 5-8% drift | The session is above the prescribed intensity by physiology, even if the watch says otherwise |
| Hot or humid environment | +2-4% drift baseline | Thermal strain is the dominant driver, not fitness |
| MAF / strict aerobic-base training | Goal: keep drift < 5% across multi-month blocks | Improving aerobic base shows up as drift dropping at matched pace |
Known Limitations
- •Drift is only interpretable on steady-pace efforts at sub-threshold intensity. Intervals, hills, surges, and tempo work all produce drift signals that mean different things or no useful thing at all.
- •Environmental conditions dominate. A 5% drift on a 32°C humid day is normal physiology; the same 5% drift on a cool 12°C day at the same pace is a fitness or readiness signal. Without ambient temperature in the model, drift is hard to compare across sessions.
- •Hydration status, caffeine, recent meals, and sleep all bias drift. A single session's drift number is noisy, the multi-session trend at matched intensity and conditions is what's actually informative.
- •Drift requires at least ~45-60 minutes of steady work to be meaningful. Short sessions don't develop enough thermal and cardiovascular strain to produce a stable signal.
Science Context
Cardiac drift was first characterized by Coyle and González-Alonso (2001) as a consequence of progressive dehydration and rising core temperature reducing stroke volume, the heart compensates by increasing rate to maintain cardiac output. The Pa:Hr (or Pw:Hr) decoupling expression, popularized by Friel and TrainingPeaks, reframes drift as a ratio so it's directionally consistent across run/bike/row. Lambert et al. (2008) and Achten & Jeukendrup (2003) established the < 5% benchmark for well-matched aerobic work; values above that on a controlled-condition session are taken as a sign the prescribed intensity has slipped above the second ventilatory threshold.