Here's the silent mistake. Your watch reads 142 bpm. The app calls it aerobic. The plan calls it easy. And underneath all of that, your body is producing lactate it shouldn't be producing, your respiratory exchange ratio is climbing, and the adaptation Zone 2 is supposed to deliver — mitochondria, fat oxidation, capillary density — is being quietly hijacked.

You did everything the screen told you. The physiology didn't agree.

This is the most common, most expensive, and least visible mistake in serious-amateur cycling. Most cyclists are training Zone 2 about ten to fifteen beats too high — not because they're sloppy, but because the labels on their training apps and the lactate response of their bodies are not the same thing. This piece is about why that happens, what the data actually says, and how to fix it without a $400 lactate meter.

Section 01What Seiler's data actually shows

In 2010, Stephen Seiler published the paper that became the foundation of modern endurance training (Seiler 2010). His lab at the University of Agder in Norway spent the previous decade cataloguing what world-champion endurance athletes were actually doing in training — not what coaches said in interviews, but the real heart-rate files, the real session logs. Rowers, cross-country skiers, distance runners, road cyclists.

The pattern has been replicated across sports and across decades. Tønnessen and colleagues found a similar low/high distribution across the annual training of 8 world-champion orienteers — with phase-specific shifts in the exact ratio across base / build / competition periods, but the overall sub-LT1 dominance held (Tønnessen et al. 2015). Lucía's earlier work on professional road cyclists at the Tour de France showed the same shape — even at the absolute top of the sport, the vast majority of riding sits well sub-threshold (Lucía et al. 1999).

But the headline gets repeated so much it starts to feel like a slogan. Three things from the underlying data deserve more airtime than they get.

One — the volume of true low-intensity work is enormous. Elite cross-country skiers accumulate 600–900 training hours per year, and ~700 of those are unambiguously easy. The 80/20 isn't a ratio elites tap into for a week a month. It's a way of life across a season.

Two — the 20% is genuinely hard. People hear polarised and assume the hard sessions are a bit above threshold. They're not. They're VO2max intervals at 90–95% of HRmax. They aren't sweet-spot wearing a costume.

Three — the middle is empty. Stöggl and Sperlich took this directly to a controlled trial (Stöggl & Sperlich 2014). Polarised showed the largest gains across every measured variable — with statistically superior VO2max improvements over the other three groups.

Section 02How most amateurs get it wrong

The structural reason is the most boring and the most important: the labels are wrong. Open TrainingPeaks. Its Zone 2 spans roughly 56–75% of FTP. A trained cyclist sitting at the top of that range — at the top of TrainingPeaks's Zone 2 — is sitting roughly at their first lactate threshold. LT1. The exact ceiling Seiler's data says world-class athletes spend most of their time below. Not at. Below.

The label is unchanged across that range. The physiology is not.

And it feels strong. Within that band, effort scales sub-linearly with power. Pushing 5% more watts feels like maybe 2% more effort. So riders gravitate to the top of Zone 2 because the body is rewarding them for being there. Stack that across three weeks and you've drifted out of polarised into the dead middle.

The fix is to find your own LT1 and use it as a ceiling, not a target.

Section 03The four-step protocol

There's a cheap way to find that ceiling. No lactate meter, no metabolic cart, just four signals you can already read.

1. The talk test, strictly.

   Four-sentence conversation without breath catching = below LT1. Single sentence breaks across two breaths = above. Every 10 minutes on new-block rides.
2. Cardiac drift < 5 bpm.

   Hold flat wattage 60 min after warm-up. HR climbing >5 bpm without wattage/terrain/temp change = above LT1 (Sylta et al. 2016).
3. Next-morning resting HR.

   A truly aerobic ride leaves morning HR within 2–3 beats of baseline. A 5+ bpm jump = prior day was glycolytic, not aerobic.
4. RMSSD-trend confirmation.

   7-day rolling RMSSD flat or rising across a 3-ride Z2 block = intensity was honest. Drifting down = too hard (Vesterinen et al. 2016).

Four signals at four different time horizons, all triangulating on the same question: was that ride actually aerobic? No one signal is enough. The four together are very hard to fool.

Section 04Apply it with HELIOS

HELIOS samples HR every two seconds at the finger — less susceptible to motion artefact than wrist optical sensors (Bent et al. 2020). ULTRA holds the 7-day RMSSD baseline, the 30-day cardiac-drift trend at matched wattage, and last night's resting HR. The daily call isn't an opinion — it's the four-signal protocol, automated.

Section 05Where this gets misread

Section 06The four-week build

The volume scales with what your baseline supports. The distribution stays constant.

Section 07Counterpoint: the volume floor

Seiler's data is drawn from athletes training 14+ hours per week. At that volume, 80% of training is, in absolute terms, a lot of low-intensity work — 11–12 hours of sub-LT1 riding is a meaningful adaptive stimulus by itself.

If you're training less than six hours per week, the math changes. Sylta, Tønnessen and Seiler (Sylta et al. 2016) found that at lower training volume, sweet-spot and threshold-emphasis blocks produced equivalent or larger gains than a strictly polarised distribution. 80% of 5 hours is 4 hours — not enough adaptive stimulus on its own.

If you're a 6-hours-a-week rider, the 80/20 isn't gospel. It's directional, not prescriptive. The honest read of the literature is that prescription should follow volume.

Section 08Bottom line

Zone 2 isn't a number on a screen. It's a physiological state, and most amateur cyclists are training it 10–15 beats too high because the label on the training app and the lactate response of the body are not the same thing.

The fix is the four-step protocol. Use your own ceiling, not the table that came with the app.

The discipline of polarised training is not in the hard work. It's in keeping the easy work easy.