Which training model works best? The one you stick to

If you have spent more time arguing polarized vs pyramidal than riding the plan on your calendar, this episode is for you.

On Escape Collective’s Performance Process, Ronan Mc Laughlin talks with Jem Arnold — physiotherapist, endurance coach, and co-author of a network meta-analysis on training intensity distribution (TID). The research question sounds like social-media bait: which model improves VO2max and time-trial performance most? The honest answer is closer to “yes” — several coherent models work when athletes actually do them.

Listen: Escape Collective episode · Apple Podcasts

This article translates that conversation into a practical N+One framework: stop treating model labels as magic, protect progression, and use connected training data to decide the next session.

What the study actually compared

Arnold’s co-authored review pooled individual-participant data across endurance training trials that quantified time in intensity domains and reported VO2max or time-trial outcomes (PubMed).

In plain language, the included work compared common TID labels using a three-zone lens (low / moderate / high), not a marketing five-zone spreadsheet:

• Polarized: mostly easy (zone 1), some hard (zone 3), very little moderate (zone 2)
• Pyramidal: mostly easy, then moderate, least time hard
• Threshold-heavy: more time around tempo / sweet spot / threshold
• Plus low-only and high-only extremes in the network

Thirteen studies and roughly 350 athletes made the cut. Volume was handled carefully so “more hours” did not masquerade as “better model.” Female athletes were still underrepresented — a real limitation when you generalize.

The headline result

Overall, polarized did not clearly beat pyramidal (or the other included models) for VO2max or time-trial performance. That does not mean intensity distribution is irrelevant. It means within structured endurance training, model branding is a weak lever compared with doing progressive work you can recover from.

A useful nuance from subgroup signals: competitive athletes may tilt toward polarized for VO2max responses, while more recreational athletes may respond well to pyramidal. That matches lived coaching: pros often look “polarized” because huge easy volume makes hard sessions sparse by necessity; time-crunched amateurs who copy pro percentages without the volume often invent gray-zone weeks instead.

Why the internet still treats models like religions

Three traps keep the debate louder than the data:

1. Survivorship stories. One model “worked” for someone who also slept, fueled, and trained consistently for years.
2. Pro cosplay. Copying elite zone shares without elite hours turns polarized into accidental tempo creep.
3. Percentage obsession. Weekly pie charts become the goal. Adaptation does not care about your screenshot.

In N+One terms: the label is secondary. The job of the week is one clear quality stimulus, enough easy volume to support it, and recovery that lets you repeat the pattern.

What to do this week instead of re-picking a model

1. Name the distribution you are already running

Look at the last 2–3 weeks of rides. Are hard days truly hard and easy days truly easy (zone 2 guardrails)? Or did “endurance” become sneaky tempo (tempo creep)?

If your week is already mostly easy with one or two purposeful hard sessions, you are closer to polarized. If you keep a moderate sweet-spot / tempo diet under a big easy base, you are closer to pyramidal. Either can work — the how-to guide covers application detail.

2. Protect progression, not purity

Pick one primary quality session for the next 14 days (threshold, VO2, or race-specific). Everything else supports it. Do not rebuild the entire TID philosophy because one interval set felt flat.