Polarized vs pyramidal training: which is better?
Neither wins. Two meta-analyses found no difference in race or time-trial performance between polarized and pyramidal training. What both have in common matters more: most of the running is easy.
What to do
- Keep most of your running easy, roughly 75–85% of your running time.
- Pick the model you can stick to. Consistency over months beats the label on the plan.
- If you are a recreational runner, a pyramidal week (lots of easy, some steady or threshold work, a little very hard work) is a sensible default.
- Don't buy a plan or an app because it promises that 80/20 polarized is proven better. It isn't.
- Avoid making most of your training moderately hard. Threshold-heavy training did worse than both models.
What the two models are
Both models split training into three intensity zones, usually set by the two lactate or breathing thresholds:
- Zone 1: easy. You can hold a full conversation.
- Zone 2: between the thresholds. Steady, “comfortably hard” running, such as marathon pace for many runners or threshold runs.
- Zone 3: above the second threshold. Hard intervals that you can only hold for minutes.
Polarized training puts about 80% of training in zone 1, little or nothing in zone 2, and the rest, roughly 15–20%, in zone 3. This is the model behind the “80/20” books.
Pyramidal training also puts most training in zone 1, but has more in zone 2 than in zone 3. Lots of easy, some steady, a little very hard.
The difference between them is smaller than the debate suggests. Both are mostly easy running.
What the meta-analyses found
A meta-analysis pools several studies to get a clearer answer than any single one. Two recent ones looked at exactly this question.
Oliveira and colleagues (2024) pooled 14 studies with 437 participants. Polarized training gave a small advantage in VO2peak, the maximal oxygen uptake. But there was no difference in time-trial performance, which is what actually matters on race day. The VO2peak advantage showed up only in programmes shorter than 12 weeks and in highly trained athletes.
Rosenblat, Seiler and colleagues (2025) went a step further. They pooled the raw data for each individual participant from 13 studies (348 people). They found no difference for VO2max or for time trials. They did find a hint that the answer depends on level: recreational athletes tended to do slightly better on pyramidal training and competitive athletes on polarized training. Neither of those subgroup effects was statistically significant on its own.
A third review of 20 studies (Campos and colleagues, 2022) found the common thread: at least 70% of training at low intensity. Both polarized and pyramidal training did better than training built mostly around threshold work.
The study the 80/20 books cite
Much of the case for polarized training rests on a 2014 study of recreational 10 km runners by Muñoz and colleagues. It is often quoted as showing a 5% improvement with polarized training against 3.6% with a more threshold-heavy plan. The study itself found no significant difference between the groups overall. Only the runners who followed their plan most closely showed an advantage for polarized training.
That doesn’t make the 80/20 books wrong about the main message. Mostly easy running is well supported. The claim that polarized beats pyramidal is not.
What real marathoners do
Trials can only run for weeks or months. Training logs from real runners show what happens over a whole marathon build.
In a study of 119,452 marathoners and more than 150,000 races (Muniz-Pumares and colleagues, 2025), the fastest runners ran more than three times the volume of slower runners in the 16 weeks before the race. Their time in the harder zones was about the same as everyone else’s. Almost all of the extra was easy running. More than 80% of these fast runners trained in a pyramidal pattern.
Elite runners look the same: at least 80% of their running is easy all year, and marathoners are more pyramidal than middle-distance runners (Haugen and colleagues, 2022).
Order may matter more than the model
One trial suggests that how you sequence the models matters. In 60 well-trained amateur runners, those who trained pyramidally first and then switched to polarized training for the final phase improved the most, including 1.5% faster over 5 km (Filipas and colleagues, 2022). This is a single trial, so treat it as an idea worth trying, not a rule.
How you count changes the label
Whether a plan looks polarized or pyramidal can depend on how you count. Applied to the same world-class training logs, counting by race-pace zones and counting by physiological zones gave zone-3 shares of 4.1% and 6.6% (Kenneally and colleagues, 2021). Counting by time or by number of sessions also gives different answers. Two coaches can describe the same week with different labels.
What we don’t know
- No trial has used marathon performance as its outcome. The trials measure time trials, usually much shorter than a marathon.
- Most trials are short, often 8–16 weeks.
- Women made up only about 15% of the participants in the pooled data.
- The observational data can’t separate cause from selection. Fitter runners may simply choose to run more.
Sources
- Oliveira et al. (2024). Meta-analysis of polarized vs other intensity distributions.
- Rosenblat et al. (2025). Individual-participant network meta-analysis of training intensity distribution.
- Campos et al. (2022). Review of training intensity distribution in endurance athletes.
- Muñoz et al. (2014). Polarized vs threshold training in recreational runners.
- Muniz-Pumares et al. (2025). Training characteristics of 119,452 marathon runners.
- Haugen et al. (2022). The training characteristics of world-class distance runners.
- Filipas et al. (2022). Effects of pyramidal and polarized sequencing in well-trained runners.
- Kenneally et al. (2021). Training intensity distribution analysis by race pace vs physiological approach.
Last reviewed 12 November 2026.