What does the review add?
Stöggl and Sperlich’s 2015 focused review brings together retrospective training records and prospective interventions across endurance sports. Its central tension is useful: many observed elite-athlete distributions were pyramidal, while several intervention studies favoured polarized programs over the alternatives they tested. Those findings do not have to contradict one another. They come from different populations, seasons, designs, and counting methods. [1]
The review does not identify a single optimal distribution for everyone. How the training was counted often explains why the reported mixes differ.
What kind of paper is this?
This is a focused review, not a new randomized experiment and not a pooled meta-analysis with one numerical treatment effect. The authors organize evidence from retrospective analyses, longitudinal and single-case studies, and randomized controlled trials. The athletes range across endurance disciplines and performance levels, and the observations cover different lengths of time.
Those categories answer different questions. A training diary can describe what successful athletes did. An intervention can compare what happened when athletes followed different programs. A single-athlete case can be richly detailed but cannot establish that another athlete would respond the same way.
Two ways to be mostly easy
Polarized
Easy dominates. Hard exceeds moderate.
Pyramidal
Easy dominates. Moderate exceeds hard.
A polarized pattern emphasizes the low and high ends with relatively little middle-zone work. A pyramidal pattern has a descending share from low to moderate to high intensity. The review reports that a large amount of low-intensity training is common to both. The distinguishing feature is the relationship between the smaller moderate and hard components.
This matters because an easy-dominant chart containing some purple does not automatically contradict the endurance literature. The review documents meaningful threshold work in some disciplines and phases. Calling all middle-zone work “junk” would turn a useful description of training into a much broader claim than the evidence supports.
Sport and season change the picture
The paper separates preparation, pre-competition, competition, and longer-term records. That organization is a reminder that a training distribution is a snapshot over a chosen interval. A block preparing for a race may not resemble a recovery period or the average of an entire year.
The review describes differences among runners, cyclists, rowers, skiers, swimmers, and other endurance athletes. Those differences can reflect the demands of the sport, how training is organized, and how its intensity is measured. They should not be flattened into a universal template for a recreational runner’s week.
For an app, the practical implication is to preserve access to different time windows. A daily view can explain an interval session. A weekly view can show how sessions are combined. A longer view can expose the cumulative pattern, while also smoothing over deliberate changes within it. Each view removes some detail and reveals something else.
Measurement can change the apparent distribution
The review discusses internal measures such as heart rate and perceived effort, external measures such as speed or power, and session-goal approaches. These are not interchangeable descriptions of the same thing. Heart-rate time is distributed within a workout; a session goal can place an entire workout into a category.
One example discussed in the review reports elite Norwegian skiers and biathletes at about 91% low-intensity time versus 9% in higher zones, but about 77% versus 23% using a session-goal approach. Both figures describe the same training.
Our interactive example uses a simpler hypothetical week to show the same counting issue. Four easy sessions plus one interval session can be 80/20 by sessions while containing a much smaller high-intensity share of minutes.
Observation and intervention need different interpretations
| Evidence | Useful question | Important limit |
|---|---|---|
| Retrospective records | What training patterns did these athletes follow? | Success does not prove that the observed pattern caused it. |
| Single cases / uncontrolled changes | What happened around a change in this athlete or group? | Other changes can contribute to the outcome. |
| Controlled interventions | How did these programs compare during the study? | Participants, duration, total dose, and tests bound the conclusion. |
The review’s intervention evidence includes the authors’ 2014 trial. That is a relevant source, but it is not an independent replication simply because it appears again in a review. Counting papers without tracing their underlying participants can make a body of evidence look larger than it is.
Where this review leaves uncertainty
Training definitions and methods vary across the underlying studies. Long-term experimental comparisons are limited, and what works during one part of a season may not describe an entire year. The review explicitly concludes that an optimal distribution cannot be identified from the available evidence.
Published in 2015, the review gives a useful starting point for understanding the different methods. The 2024 paper notes extend the reading with pooled estimates, confidence intervals, and a link to a 2025 update.
When a week has more moderate training than usual, open the workouts behind it. It might be a planned threshold session, a hillier route, or an easy run that got away from you. The total alone won’t tell you which.
Sources
- Stöggl TL, Sperlich B (2015). The training intensity distribution among well-trained and elite endurance athletes. Frontiers in Physiology 6:295. DOI: 10.3389/fphys.2015.00295.
- Original article © 2015 Stöggl and Sperlich, published under Creative Commons Attribution. Summary adapted by Zonometry.