The cycling internet has a favourite training model, and it fits on a bumper sticker. Eighty percent easy, twenty percent hard. Job done, go ride.
The problem is that almost every podcast, YouTube video, and Zwift coach repeating this line has stripped out the parts that make it useful. Prof. Stephen Seiler — the researcher whose name is attached to the model — has spent a decade patiently correcting the same misunderstandings, and the corrections aren't reaching the people who need them.
This piece takes a stance. The popular version of polarised training is wrong enough to hurt your results, and the fix isn't complicated. It just requires reading past the headline.
The oversimplified version: '80/20 and done'
Scroll through cycling content for ten minutes and you'll hear the same pitch. Do 80% of your riding in Zone 2. Do 20% hard. Avoid the middle. That's polarised training.
It's tidy, it's memorable, and it's wrong in several specific ways. First, the 80/20 split Seiler described is a count of training sessions, not a percentage of time or TSS. An athlete doing five Zone 2 rides and one interval session per week is running roughly an 83/17 session split even if the interval day is only 90 minutes against 20 hours of easy volume.
Second, the popular version treats "hard" as anything uncomfortable. Sweet spot, threshold, and VO2max all get lumped together. Seiler's definition of the hard 20% is specifically work above the second lactate threshold — typically VO2max intervals at 105-120% of FTP, not 30-minute threshold blocks.
Third, and this is the one that matters most: the model was derived from observing what elite rowers, cross-country skiers, and runners actually do over a season. It's descriptive. The cycling internet has converted it into a prescription for a 42-year-old amateur doing eight hours a week, which is not the same problem.
The result is thousands of riders grinding out Zone 2 five days a week, then doing a tempo ride on Saturday, calling it polarised, and wondering why their FTP won't move.
What Prof. Seiler actually said
Seiler's original papers and his subsequent interviews — including long-form conversations on Prof. Stephen Seiler's guest page — are more careful than the meme version.
His central observation is this: across multiple endurance sports, elite athletes who sustain high training volumes converge on a similar intensity distribution. Roughly 80% of sessions sit below the first lactate threshold (LT1, around 65-75% of FTP for most cyclists). Roughly 20% sit above the second lactate threshold (LT2, approximating FTP). Very few sessions sit in the middle zone.
Seiler has said repeatedly that this is not a moral claim. He does not argue every athlete should train this way. He has acknowledged in multiple interviews that pyramidal distributions — where threshold work features more prominently — produce excellent results, particularly in build phases and for events where threshold power is the specific determinant.
What he does argue, consistently, is that the easy days must be easy. His phrase is "intensity discipline". The failure mode he has observed in amateurs is not too little hard work. It's too much moderate work, which blunts recovery before the hard sessions and never loads the low-intensity adaptations properly.
He also stresses that the 20% hard component should include genuine VO2max stimulus — 4x4 minute intervals, 8x4s, 30/15s — not endless sweet spot. If your hard sessions top out at 95% of FTP, you are not training polarised. You are training pyramidal and calling it something else.
Why the 80/20 rule is a description, not a prescription
This distinction gets lost constantly. A description tells you what successful athletes happen to do. A prescription tells you what you should do. They are not the same thing.
Seiler's data came from athletes training 15-30 hours per week. At that volume, an 80/20 session split produces a staggering amount of aerobic base — 12+ hours a week of Zone 2 for a 15-hour athlete. The adaptations stack up because the absolute dose is enormous.
An amateur riding eight hours a week cannot replicate that dose. Eighty percent of eight hours is 6.4 hours of Zone 2 — about a third of what the elite model provides. If that rider then copies the elite 20% hard ratio, they get 1.6 hours of high intensity, which is also under-dosed relative to what most amateur physiologies respond to.
This is why Dan Lorang, former head of performance at Red Bull-Bora-Hansgrohe and long-time coach to Jan Frodeno and Gustav Iden, has said repeatedly that amateur athletes often need proportionally more intensity than pros, not less. Their weekly volume ceiling is lower, so the threshold and VO2max work has to carry more of the adaptive load.
The deeper point: a distribution that works for a 25-hour pro is not automatically right for an 8-hour amateur. Copying the shape of elite training without the volume underneath is one of the most common polarised training mistakes we see in riders coming to us from self-coached plans.
The three things the 80/20 frame leaves out
First, periodisation. The elite athletes Seiler studied don't train 80/20 every week of the year. Their distributions shift across base, build, peak, and race phases. Early base work is often more pyramidal, with tempo and threshold featuring heavily. Peak phases polarise sharply. The annual average lands near 80/20, but no single week looks like the average.
Second, session structure. The 80/20 headline says nothing about how the hard 20% is delivered. Four VO2max sessions per week at 4x4 minutes is a very different stimulus from two sessions of 8x4s, which is different again from one long threshold day plus one VO2 day. Seiler's work suggests frequency of hard sessions matters — 2-3 per week is the sweet spot for most trained athletes — but the popular model is silent on this.
Third, the athlete. Time-trialists living at threshold for 40 minutes need different preparation than criterium riders who live above threshold in 30-second bursts. Triathletes have a run to protect and can't absorb the same bike intensity as a pure cyclist. A rider coming back from illness needs a different distribution than one peaking for a target event.
This is why we don't sell a polarised plan through our coaching. We sell a periodised plan that uses polarised distributions when they serve the athlete and pyramidal distributions when those fit better. The frame matters less than the execution.
How to structure a week that honours the research
If you're training 8-12 hours per week and want to apply what Seiler's work actually supports, here's a defensible template.
Three to four Zone 2 rides per week, strictly capped below LT1. For most riders that's 65-75% of FTP, conversational, nose-breathing possible for most of the ride. If you finish these rides feeling like you pushed, they were too hard. That's the discipline Seiler keeps emphasising.
Two hard sessions per week, spaced 72 hours apart. One VO2max session — 4x4 minutes at 110-115% of FTP, or 5x3 minutes, or 30/15s for 20-30 minutes of work — and one threshold or over-under session of 2x20 minutes at 95-100% of FTP. This gives you a pyramidal-leaning week during build phases and a cleaner polarised week if you drop the threshold day during a peak block.
One longer endurance ride on the weekend. This should be your biggest aerobic dose — 3-5 hours at Zone 2 with maybe one 20-minute tempo block if the event you're preparing for demands it. Don't turn the long ride into a smash-fest because it feels productive.
One full rest day. Not active recovery, not "just spinning". Off the bike.
That's the week. It respects the session-count ratio, delivers genuine high-intensity stimulus, and protects the easy days from drift. It also flexes — add a second long ride in base, drop the threshold session in peak week, shift the VO2 session earlier if racing Saturday.
Start this week. Pick your two hard days, cap your easy days at 75% of FTP, and run the block for four weeks before you judge it.
