Field note
Durability is how well your pace and heart-rate cost hold up in the second half of a long effort, compared with the first half.
Fitness often shows up late, not early. Splitting one steady outdoor run into equal-distance halves and comparing speed per heartbeat in each half is a transparent field proxy related to that idea. It is not a lab measurement and not a fitness verdict by itself.
Mean speed in each half is calibrated GPS distance divided by pause-excluded active time. Mean heart rate comes only from measured samples inside that half. A workout average is never substituted.
For each half, divide mean speed by mean measured heart rate. The displayed percentage is 100 times one minus the back-half ratio divided by the front-half ratio. Positive means the back-half ratio was lower. Negative means it was higher. The two half values have units. Only their ratio and the displayed percentage are unitless.
| Sign | Plain reading |
|---|---|
| Positive % | Back half cost more heart rate for the speed you held |
| Negative % | Back half looked better on speed per heartbeat |
| Near zero | Front and back halves stayed similar |
Maunder and colleagues (2021) define durability more broadly as the onset and magnitude of deterioration during prolonged exercise. Pace-heart-rate drift is one field proxy related to that wider idea. It does not measure durability in the lab, aerobic fitness, or training improvement by itself.
Coyle and González-Alonso (2001) describe cardiovascular drift: heart rate can rise during sustained work even when external pace stays similar. Heaps and colleagues (1994) showed hypohydration can drive that drift without reducing blood volume. Heat, hydration, grade, wind, pacing, heart-rate lag, and sensor gaps can all move the comparison. One workout cannot identify a cause.
TrainingPeaks (2023) uses less than five percent after at least twenty minutes as a coaching reference for aerobic decoupling. That is a field convention, not a clinical cutoff, and it should not become a green, amber, or red verdict.
Compare similar steady runs over time if you use this number at all. Keep route, conditions, duration, and pacing as alike as practical. Never treat one workout as a verdict. This sits beside What is VO2max on a watch? for the longer fitness estimate and beside What is cardio load? for how much heart work the whole session cost.
One related field number is efficiency factor for the whole run: mean speed in metres per minute divided by mean measured heart rate. A higher number means more ground for the same heart cost. Read it against your own comparable runs. It moves with heat, hills, and pacing just like the durability read. TrainingPeaks (2023) documents Efficiency Factor as a coaching tool; official running EF uses normalized graded speed, which many consumer watches do not calculate.
Relative intensity asks how hard this session was for you next to your own recent runs of the same kind. Banister (1991) framed modeling athletic load from heart-rate cost over time. One means a typical session. Below one is easier than usual. Above one is harder. It needs several comparable recent runs before it means much, and a single session is always noisy.
How fast you settle afterward is a different question. That lives in How fast should my heart settle after exercise?
Stay quiet unless the run has at least twenty minutes of pause-excluded route time, measured heart-rate coverage in both halves, and a steady enough pace. Steadiness gates are conservative calibrations, not scientific thresholds.
Hills can be corrected toward flat-ground equivalent using grade-adjusted energy cost, so a climb in the second half no longer looks like fading by itself. On very steep ground that correction is least trustworthy, because lab treadmill models give back more on descents than real legs do. Past that steepness, saying nothing is better than publishing a number the correction cannot carry.
Normalized Power and Variability Index still need a power meter. Heart-rate field versions are honest within-person trends, not lab measures.
Fitness often shows in the second half. One percentage is a clue, not a grade. That is the line worth remembering.
I write these because I build Aera, an iPhone app that reads Apple Health against your own baseline and puts the research and the error bars underneath every number it shows you, including this one. You do not need it to use anything above.