Field note
Interval workouts are structured sessions made of efforts and recoveries, each kept as its own step rather than melted into kilometres.
A 3000 does not end where a kilometre ends. If the watch recorded the session as steps, the honest read is those steps. Each effort and each recovery stays its own thing. Where the watch has no boundary, kilometre splits keep the page. That is the honest outcome, not a missing feature.
Two things arrive from the watch. The first is the recipe, the session as you built it. It knows a step was meant to be a 3000 and the next one was meant to be a rest. The second is the recording, a mark laid down each time a step actually ended. Line the two up. The recording decides where a step starts and stops, because that is what happened. The recipe only supplies the name.
Where they disagree, the recording wins and the name goes. A rep that came in well under the distance it was meant to be keeps every one of its numbers and quietly loses the label. Calling it a 3000 would be the one dishonest thing on the page. If too few steps line up to trust the recipe at all, the names are dropped together. The steps are simply numbered.
| Source | What it is honest about |
|---|---|
| Recipe | What you meant each step to be |
| Recording | Where a step actually started and stopped |
| When they disagree | Keep the numbers. Drop the label. |
| No boundaries at all | Stay with kilometre splits and say so |
A session run without a recipe, or one where you never pressed lap, has no steps to find. A watch that cuts the run automatically at every kilometre is not describing a session. That pattern is a distance split, not a workout plan.
It is tempting to guess where a rep began from the heart-rate trace. That is the wrong shortcut. Wrist heart rate is at its least accurate exactly where a rep boundary lives. Error roughly doubles while your heart rate is changing quickly, compared with steady running. Van Oost and colleagues (2025) measured wrist-worn photoplethysmography during those rapid transitions. That is the moment a lap button, not a pulse, has to do the work.
Your heart also lags the effort by the better part of a minute. A rep guessed from heart rate would start late and end late. Every number hanging off it would be wrong by the same amount. Coaches control intervals by pace and never by pulse for that reason.
So when the watch recorded no boundaries, there is no ladder of reps to invent. Heart-rate zones can still describe the session as a whole. They cannot invent the edges of a 3000. That distinction sits beside What are heart rate zones?
An effort and a recovery are not two versions of the same thing. Buchheit and Laursen (2013) treated high-intensity interval training as a programming puzzle with separate levers. The work bout sets the stimulus. The recovery decides how much of that stimulus you carry into the next rep. Ranking your jog against your rep is a category error. They should not be drawn in one grammar.
A useful field picture keeps them apart. A rep is a figure you can compare with the other reps. A recovery is a quiet line with its time and its heart rate, and no ranking at all. The reps compare only with each other. The scale stretches to the spread of that set. A session that held together draws as a row of near-equal efforts. A session that came apart draws as a staircase. The comparison is not your pace. It is how far this rep sat from the rest of them.
Abbiss and Laursen (2008) reviewed how athletes spread effort across a competition. The same idea applies inside a set. Holding is not automatically better than fading. A set that fades on purpose is a session too. How you spread work across a continuous run is a sibling question, not the same one. That lives in What is pacing?
The figure on each rep is its time, because a step is a thing that took time. Pace rides beside it. Average heart rate sits with them. Power and cadence, when the watch recorded them for that rep alone, belong to that step and not to the kilometre they happened to fall inside.
Everything here rests on what the watch wrote down. Distances come from GPS and drift a little, so a rep may read a few metres off what you actually ran. Times and heart rates come from the recording. They are as good as the watch that made them.
The tolerances that decide when a name is trusted, when a rep is too short to keep its label, and when a boundary is a stray press rather than a step, are field judgements. Nobody publishes a recipe for reading a watch session back. These are calibrations, not scientific thresholds.
When there is no GPS track at all, like lifting or a circuit, the honest question changes. You are no longer reading steps. You are asking whether the heart trace itself rose and fell. That lives in What is work and rest rhythm?
Read the session as the steps you ran. Infer nothing the watch did not mark. 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.