Are You a Polarized or Pyramidal Responder? What 2025 Research Says About Your Training Type

Your training partner swears by 80/20. You got fitter on threshold work. Both can be true. The best 2025 data found no average winner between polarized and pyramidal plans, which is exactly what you’d expect if different people answer to different styles.

Polarized training puts most hours very easy, a little very hard, and almost nothing in the middle. Pyramidal training keeps most hours easy, a good chunk at tempo, and a little at the top. The general debate is covered in our polarized vs pyramidal breakdown. This post asks a different question: which one fits you?

What the polarized vs pyramidal responder research found

The strongest 2025 paper pooled data from 13 studies and 348 athletes. The effect on VO2max (your body’s peak oxygen use) was -0.06. For time-trial performance it was -0.05. Both are basically zero.

In short: on average, the two styles tie.

An older 2024 analysis of 14 studies found a small edge for polarized on VO2peak (0.24) and none for time trials (-0.01). That edge showed up mostly in short blocks under 12 weeks and in highly trained athletes.

Look closer at that edge. The VO2peak result came from 11 studies and 284 athletes. Its confidence range ran from 0.01 to 0.48, so the low end barely clears zero. The time-trial result came from 221 athletes and showed nothing. Translation: a small lab gain didn’t turn into faster racing.

Now look at what happens inside one real trial. In 60 well-trained runners over 16 weeks, the average gains looked like this.

Two things stand out. The groups that switched styles gained the most. And the person-to-person spread was as big as the gain. Averages hide people.

The race results told a similar story. Five-kilometer time dropped 0.6% on pyramidal and 1.1% on polarized. It dropped 1.5% for runners who moved from pyramidal to polarized, and 0.9% for the reverse. The authors also noted that most individual changes were smaller than the smallest change their tests could reliably detect. In short: the order you train in may matter, and many single-athlete changes are too small to trust.

Why a “no difference” average can hide two kinds of athletes

Picture a classroom thermostat. Half the room runs hot, half runs cold. The average temperature is perfect, and nobody is comfortable.

A zero average works the same way. If one half of athletes gains from polarized and the other half from pyramidal, the pooled effect lands near zero. The authors of the 2025 analysis didn’t claim this. It’s our reading of the numbers, and it fits the pattern.

Another clue comes from training level. The pooled analysis found that VO2max response differed between recreational and competitive athletes. Summaries suggest competitive athletes lean polarized and recreational athletes lean pyramidal. Treat that as a hint, not a rule. It fits what we found about intensity blindness: at under 10 hours a week, strict 80/20 often has no real edge.

One 2025 machine-learning trial: read it with caution

A Korean team ran a 16-week trial with 120 marathoners, split evenly between the two styles. They fed 27 measures into machine-learning models and sorted athletes into four groups: polarized responders (31.5%), pyramidal responders (31.9%), responders to both (18.7%), and non-responders (17.9%).

It’s a fun idea. Treat it as a hint, not a fact. It’s one center, one study, no replication, and it has small groups and numbers that don’t add up cleanly. The percentages don’t match whole athletes. Nobody should read those slices as the share of the population.

The useful takeaway is simpler. Response probably varies by person.

How much “non-response” is just noisy testing?

Here’s the catch with any self-test. Your numbers wobble even when nothing changes.

A crossover study with 21 people defined non-response as a change smaller than twice the typical error of the test. For VO2peak that was 0.107 L/min. For lactate threshold it was 15.7 W. For heart rate at a fixed effort it was 10.7 bpm. Every person in that study responded to at least one measure.

Translation: a “no change” result might just be a blurry ruler.

Heart rate at a fixed effort is shaky. A hot day or a poor night’s sleep can move it by more than a training block does. So don’t judge a block on one run.

A 6-week self-test with two field markers

Researchers haven’t yet shown that these field markers can pick your training type. Treat this as a personal experiment, not a diagnosis.

Pick two markers you can repeat:

• Pace at LT1. LT1 is your first lactate threshold, the top of true easy running. Note the pace you hold at that effort. See Zone 2 vs LT1 for finding it.
• Drift ratio. This is the change in your heart-rate-to-pace efficiency between the first and second half of a steady run. Pros call it Pa:Hr. A lower number is better. Our guide to aerobic decoupling explains the benchmarks.

Then follow a simple plan:

1. Run a 6-week block leaning polarized. Keep volume steady.
2. Re-test both markers on the same course, in similar weather.
3. Run a 6-week block leaning pyramidal at the same volume.
4. Re-test and compare.

A few rules keep the test honest. Run the test at the same time of day, on the same flat route, after an easy day. Skip it when you’re sick or badly under-slept. Change one thing at a time, so don’t swap shoes, routes and plans together. If you can, repeat the test run twice, a week apart. The gap between those two runs is your personal noise. Any block-to-block change smaller than that gap tells you nothing.

Only act on a gap that is clearly bigger than your week-to-week wobble. If both blocks look the same, you probably respond to both. That’s good news. Pick the one you’ll actually stick with.

Small gaps lie. Repeated gaps don’t.

A mini case study

Take a runner I’ll call Dana. She’s 41, trains eight hours a week, and wants a 3:45 marathon. Her first 6-week block leaned polarized, with three easy runs, one hard interval day and a long run. Her drift ratio on a steady 90-minute run fell from 8.1% to 6.9%. Her pace at LT1 barely moved.

Her second block added a weekly tempo run. Drift fell from 6.9% to 4.8%, and her pace at LT1 improved by about 8 seconds per mile. Her week-to-week wobble on the same course was roughly 1 point of drift. That second block beat the noise. For Dana, tempo work earned its place.

This is one runner on one plan. Another athlete could easily see the reverse. Dana’s result doesn’t say pyramidal is better. It says Dana’s own data pointed one way.

Which pattern fits you first?

Use the research as a starting guess, not a rule.

A plan you follow beats a perfect plan you skip.

Where AthleteOS fits

AthleteOS sorts every workout into easy, steady and hard time, so you can see your real split instead of your intended one. Its session analysis tracks your pace at LT1 and your drift ratio, then compares how they moved during polarized-leaning blocks and pyramidal-leaning blocks for you. The AI coach only calls a winner when the change beats your measurement noise. It’s a personal evidence check, not a validated classifier. See pricing and the 14-day trial to try it.

Your training type isn’t a label. It’s a result you can test.