You know the moment. You've read three articles, listened to two podcasts, and opened a fourth forum thread — and you're now less sure which training model to follow than when you started. Polarised says 80/20 or you're wasting your time. Pyramidal says the pros don't actually do strict 80/20. Some bloke on Reddit says sweet spot three times a week is all you need. Everyone's got a study. Nobody agrees.
Here's the thing nobody tells you: the argument itself is the problem. Not because there isn't a right answer, but because most riders are asking the wrong question. They're asking "which model is best?" when they should be asking "which model fits my life, my training hours, and where I am in my season?"
Let me break this down properly. Three models, head to head, stripped of the tribalism — and a practical framework for choosing the one that actually moves the needle for you.
The Three Models, Defined Clearly
Before we compare anything, let's be precise. Half the arguments online happen because people are comparing things they've defined differently.
Polarised (the Seiler model)
Roughly 80% of training time below the first lactate threshold (LT1) — properly easy, conversational riding. About 20% above the second lactate threshold (LT2) — hard VO2max intervals and threshold work. Almost nothing in the middle. Stephen Seiler codified this after observing that elite endurance athletes across cycling, rowing, cross-country skiing and distance running all converged on this distribution independently. The key word is independently. Different sports, different countries, same ratio.
Pyramidal
The bulk of training is still easy (Zone 1–2), but there's a meaningful wedge of moderate-intensity work — tempo, sweet spot, sub-threshold — sitting between the easy base and the hard top. If you drew the time spent in each zone as a bar chart, it would look like a pyramid: lots at the bottom, moderate in the middle, a point at the top. This is what most World Tour training logs actually show when you analyse a full season rather than a single block.
Threshold-heavy (sweet spot dominant)
The majority of structured work sits at or near threshold — 84–105% of FTP. Sweet spot intervals, tempo blocks, threshold repeats. Easy riding exists but takes a back seat to the intensity sessions. This is what most self-coached amateurs default to, whether they realise it or not, because it feels productive and the FTP gains come quickly.
Here's where it gets really interesting: these three models aren't as different as the internet makes them sound. They sit on a spectrum of how much moderate-intensity work you include. Polarised says almost none. Pyramidal says some. Threshold-heavy says lots. The question is how much of that middle band your body can absorb and convert into fitness — and that depends on factors most people never consider.
What the Research Actually Shows
The studies that favour polarised
The headline study is Stöggl and Sperlich (2014). They randomised 48 well-trained endurance athletes into four groups — polarised, threshold-dominant, high-intensity, and high-volume — and ran them for nine weeks. The polarised group produced the largest gains in VO2max, peak power, and time to exhaustion. Not marginally. Clearly.
Esteve-Lanao's 2007 work with runners pointed the same way: athletes who spent more time at low intensity and less at moderate intensity performed better in races than those who loaded up on tempo work.
Seiler's own reviews across two decades consistently show that the 80/20 distribution isn't a coaching philosophy someone invented — it's a pattern that emerges when you let experienced athletes and coaches accumulate years of trial and error. It self-organises.
The catch nobody mentions
Here's the part that changes everything for most of us. Those studies used well-trained athletes doing 12–20+ hours per week. At 15 hours a week, 80% easy gives you 12 hours of aerobic volume. That's enough to drive serious mitochondrial and capillary adaptation. At 7 hours a week, 80% easy gives you 5.5 hours. That's a very different stimulus.
Seiler himself has been open about this. When he mapped the 80/20 distribution, he was looking at athletes who had the volume to make the easy component truly productive. Drop the volume and the maths shifts. Five and a half hours of easy riding per week, spread across three or four sessions, isn't enough to generate the aerobic adaptations that make polarised powerful. The percentage of intensity has to increase to provide sufficient stimulus — which pushes you toward pyramidal or threshold-heavy by necessity, not by choice.
The studies that favour pyramidal
Bent Ronnestad's block periodisation research with Norwegian cyclists offers a different lens. His work shows that concentrating specific intensity into focused blocks — including substantial threshold work — produces significant gains in well-trained riders. The broader picture from his group's data is that successful competitive cyclists tend to train pyramidal across a season, with polarised and threshold-focused micro-blocks embedded within it.
When researchers analysed the actual training logs of World Tour professionals across full seasons, the distribution was consistently pyramidal: high volume of easy riding, a moderate wedge of tempo and threshold work, and targeted high-intensity sessions. Strictly polarised weeks appeared during specific preparation phases, not year-round.
Dan Lorang, Head of Performance at Red Bull–Bora–Hansgrohe, has talked about this directly. His riders follow a pyramidal macro-distribution with polarised weeks embedded during specific training phases. The season plan isn't one model or the other — it's both, applied at different scales.
Where threshold-heavy fits
Threshold-dominant training has the thinnest long-term evidence but the strongest short-term results. Studies of time-crunched cyclists running sweet spot blocks show meaningful FTP gains in 4–6 week windows. The problem is what happens after. The gains plateau, the fatigue accumulates, and without periodising away from threshold, the rider stalls at 8–12 weeks.
Joe Friel made this point years ago in The Cyclist's Training Bible: threshold work is a sharpening tool, not a foundation. Use it to bring fitness to a peak, not as your year-round default. The riders who run sweet spot every Tuesday and Thursday for six months straight are the ones posting on forums about why their FTP hasn't moved in a year.
The Real Difference: Your Available Hours
Strip away the ideology and here's what actually determines which model works for you.
10+ hours per week: polarised wins
At this volume, you can accumulate 8+ hours of genuine easy riding — enough for the aerobic engine to respond meaningfully. Your two hard sessions per week (VO2max intervals, threshold work, race-specific efforts) sit on top of a deep aerobic base. This is where the Seiler model was designed to operate, and the research consistently supports it.
The 10-hour rider running strict 80/20 gets 8 hours of Zone 1 work and 2 hours of hard intervals. That 8 hours of easy volume is where the mitochondrial density builds, where capillary networks expand, where fat oxidation improves — the slow, patient adaptations that Seiler has documented across decades. Try to replicate those adaptations with 4 hours of easy riding and it simply doesn't land the same way. The dose-response relationship is real.
7–10 hours per week: pyramidal is the pragmatic choice
You've got enough volume for a solid aerobic foundation but not quite enough for pure 80/20 to express its full benefit. A pyramidal approach — mostly easy, with 2–3 structured sessions that include some threshold and some high-intensity work — gives you sufficient stimulus across the spectrum. The moderate work isn't wasted here because you need it to fill the gap that lower volume creates.
This is where Joe Friel's periodisation thinking and Seiler's distribution research actually converge. Friel has always argued that the training week should serve the event and the athlete's available hours, not a theoretical model. For the 8-hour-per-week rider targeting a gran fondo, a pyramidal week with two sweet spot sessions, one VO2max touch, and the rest easy provides a broader adaptive stimulus than strict 80/20 at the same volume.
The good news is that pyramidal still protects the easy base. You're not abandoning the aerobic foundation — you're supplementing it with moderate work that wouldn't be necessary if you had another 4–6 hours of easy volume available. The principle is the same: the easy rides must be easy, the hard sessions must be hard, and the moderate work sits in a controlled, deliberate wedge rather than spreading across every ride.
Under 7 hours per week: threshold-heavy has a case
Controversial, but honest. With 5–6 hours a week, four hours of easy riding isn't doing what it does at 12 hours. You need more of each session to count. Two or three sessions with sweet spot or threshold intervals, bookended by easy riding, will produce more measurable fitness change than five short easy rides and one interval session. The trade-off is higher fatigue cost and a need to periodise more aggressively — you can't run this distribution for 52 weeks.
The irony is that most riders with limited hours already train threshold-heavy without knowing it. The difference between doing it deliberately with proper recovery management and doing it accidentally because every ride drifts to moderate intensity is the difference between progress and a permanent plateau.
The structure matters: even in a threshold-heavy week, the warm-up and cool-down portions of interval sessions should be properly easy. A 75-minute sweet spot session might contain 30 minutes of actual sweet spot work and 45 minutes of Zone 1 riding. That's not wasted time — it's the aerobic stimulus that keeps the engine ticking underneath the intensity.
What Most Amateurs Actually Do (and Why It Fails)
Here's where it gets uncomfortable. The fixable mistake most riders make is not choosing the wrong model. It's failing to execute any model.
Look at the actual intensity distribution of a typical self-coached amateur's week. It's not polarised. It's not pyramidal. It's not threshold-heavy. It's a blob — every ride landing somewhere between moderately hard and moderately easy. The easy rides creep up because easy feels lazy. The hard sessions never hit full commitment because the rider is chronically fatigued from those not-quite-easy rides. The distribution matches none of the three models.
This is the grey zone that Seiler warns about, and it's the single most common training pattern in amateur cycling. It produces the highest fatigue-to-adaptation ratio of any distribution. Maximum tiredness, minimum progress.
Whatever model you choose, the first step is the same: make easy properly easy. Below LT1. Conversational. The kind of riding where other cyclists pass you and you let them. That discipline is more important than which percentage you assign to each zone.
The talk test is your primary gate. If you can't speak in full sentences, you've already left Zone 1. Back it up with a heart rate ceiling around the point where breathing first noticeably deepens — your working LT1 — and watch for cardiac drift across longer rides. The gap between "easy" and "comfortably moderate" is smaller than most riders think, and drifting across it four times a week is exactly how the grey zone swallows your training plan.
How Coaches Actually Apply These Models
Here's something the research papers don't capture: no serious coach runs a single model all year. The model is the ingredient, not the recipe.
Dan Lorang programmes the season for his Bora riders as a pyramidal macro-structure with polarised micro-blocks embedded at specific points. The base months are heavily polarised — lots of easy volume, minimal moderate work, targeted VO2max sessions. As the race calendar approaches, moderate work increases (climbing tempo, race-pace efforts on long rides), and the weekly distribution shifts toward pyramidal. During stage races, the distribution is whatever the race demands — nobody's counting zones at the Tour de France.
Ronnestad's block periodisation approach takes this further. Rather than spreading three intensity types across every week, he concentrates one stimulus into a 7–14 day block: a week of VO2max focus, followed by recovery, then a week of threshold focus, then recovery. Each block is temporarily "threshold-heavy" or "VO2max-heavy," but the season-level distribution remains pyramidal because the easy volume dominates across months.
For amateurs, the practical takeaway is clear: don't marry one model. Use polarised principles as your default distribution — protect the easy, commit to the hard, minimise the muddle. Then layer in pyramidal and threshold-heavy blocks when the phase and the calendar call for them. The annual plan is a sequence of distributions, not a single ratio applied 52 weeks a year.
Example Training Weeks: Each Model in Practice
Three concrete weeks for a rider with 8 hours available. Same total volume, different distributions.
Week A: Polarised (80/20)
| Day | Session | Duration | Intensity | |-----|---------|----------|-----------| | Mon | Rest | — | — | | Tue | VO2max intervals: 5×4min at 110–115% FTP / 4min easy | 1.25hr | Hard | | Wed | Easy endurance spin | 1hr | Zone 1 | | Thu | Threshold repeats: 3×12min at 95–100% FTP / 5min easy | 1.25hr | Hard | | Fri | Rest or 30min easy spin | 0–0.5hr | Zone 1 | | Sat | Long endurance ride | 3hr | Zone 1 | | Sun | Easy endurance ride | 1.5hr | Zone 1 |
Distribution: ~6 hours easy, ~2 hours hard, minimal moderate. The Saturday ride is doing disproportionate aerobic work — those steady hours below LT1 are where the mitochondrial density builds.
Week B: Pyramidal
| Day | Session | Duration | Intensity | |-----|---------|----------|-----------| | Mon | Rest | — | — | | Tue | Sweet spot intervals: 3×15min at 88–93% FTP / 5min easy | 1.25hr | Moderate-hard | | Wed | Easy endurance spin | 1hr | Zone 1 | | Thu | VO2max intervals: 4×4min at 110–115% FTP / 4min easy | 1.25hr | Hard | | Fri | Rest | — | — | | Sat | Long endurance ride with 20min tempo finish at 76–85% FTP | 2.5hr | Mostly Zone 1, moderate finish | | Sun | Easy endurance ride | 1.5hr | Zone 1 |
Distribution: ~5 hours easy, ~1.5 hours moderate (sweet spot + tempo), ~1 hour hard. The tempo finish on Saturday simulates race-day fatigue patterns — moderate work on pre-fatigued legs is a different stimulus to moderate work on fresh legs.
Week C: Threshold-Heavy
| Day | Session | Duration | Intensity | |-----|---------|----------|-----------| | Mon | Rest | — | — | | Tue | Sweet spot: 2×20min at 88–93% FTP / 5min easy | 1.25hr | Moderate-hard | | Wed | Easy endurance or rest | 0.75hr | Zone 1 | | Thu | Threshold: 2×20min at 95–100% FTP / 8min easy | 1.25hr | Hard | | Fri | Rest | — | — | | Sat | Sweet spot with VO2max finisher: 2×15min SS then 3×3min at 115% FTP | 1.75hr | Moderate-hard / Hard | | Sun | Easy endurance ride | 1.5hr | Zone 1 |
Distribution: ~2.5 hours easy, ~2.5 hours moderate, ~1.5 hours hard. This is productive for 4–6 weeks. Run it for 16 weeks without a recovery pivot and you'll stall. The fatigue cost of three intensity days per week is real, and it compounds.
How to Periodise Between Models (the Practical Answer)
The honest answer that no single-model advocate wants to hear: the most effective approach for most amateurs is rotation.
Phase 1: Polarised base (8–12 weeks)
Build the aerobic floor. Protect the easy volume ruthlessly. One or two hard sessions per week — VO2max early in the phase, threshold later. No sweet spot, no tempo blocks. This is the phase most riders cut short because it doesn't feel like enough. It is.
Seiler's point about frequency, duration, then intensity applies here. Get consistent first. Stretch the rides second. Add intensity last. This is also where the base training guide picks up in detail.
Phase 2: Pyramidal build (6–8 weeks)
Introduce structured moderate work. Sweet spot intervals, tempo efforts on the end of long rides, sustained climbing repeats. The distribution shifts from 80/20 to something closer to 70/20/10, with that 10% sitting in the moderate band. The aerobic base you built in Phase 1 is what allows this intensity to be absorbed without chronic fatigue.
This is where Ronnestad's block periodisation approach applies well. Concentrate the threshold work into two-week focused blocks with recovery weeks between them. Don't spread moderate intensity thinly across every week — cluster it, absorb it, recover from it.
Phase 3: Race-specific sharpening (3–4 weeks)
Targeted VO2max work, race-pace efforts, event-specific intervals. Volume drops 15–20%. Intensity stays high but the sessions are shorter and more specific. If you're targeting time trials, threshold repeats sit here. If you're targeting road races or sportives, over-under intervals and surge-recovery work sit here.
Phase 4: Taper and race (1–2 weeks)
Volume drops another 30–40%. Keep two short, sharp intensity touches. Everything else is easy. Trust the fitness.
Phase 5: Transition (1–2 weeks)
Unstructured. Ride when you feel like it, however you feel like it. The body adapts during rest. Skip this phase and Phase 1 of the next cycle starts on a worse foundation.
The Metrics That Tell You It's Working
Three numbers to track across any model, measured over 8–12 week blocks:
FTP trend. Not a single test — the trend across multiple tests or estimated from workout data. Rising FTP means the stimulus is working. Flat FTP after 8 weeks means the current distribution isn't producing enough adaptive stress.
Heart rate at a given power. Pick a steady-state power you can sustain for 30+ minutes and track your average heart rate at that output. If heart rate is stable or dropping at the same power, your aerobic efficiency is improving. If heart rate is climbing at the same power, fatigue is accumulating faster than adaptation.
RPE at threshold. What does 95% of FTP feel like? If a 20-minute threshold effort feels progressively more manageable across a training block, the model is working. If the same effort feels harder week on week, you're digging a fatigue hole. The FTP zone calculator sets the exact targets, but RPE is the leading indicator that tells you whether those targets are being absorbed.
If all three metrics are moving in the right direction, don't change the model. If two or more are stalling or regressing, it's time to reassess — either the distribution doesn't suit your current volume, or the execution within the model is off (usually easy rides that aren't easy enough).
Five Mistakes That Lock Riders Into the Wrong Model
1. Treating the model as an identity. "I'm a polarised athlete" is not a useful sentence. You're a cyclist who might benefit from a polarised distribution right now, given your current volume and training age. Next phase, that might change.
2. Running one model year-round. Polarised works brilliantly as a base approach. It's less effective as a sharpening approach. Threshold-heavy works brilliantly as a 4–6 week build block. It stops working if you never leave it. The models are tools, not religions.
3. Choosing a model that doesn't match your hours. If you've got 6 hours a week and you're running strict 80/20, you're getting less than 5 hours of easy riding — probably not enough volume for the aerobic component to drive meaningful change. Be honest about your time and pick accordingly.
4. Arguing about distribution while ignoring execution. The difference between a perfectly designed polarised week where easy rides sit at 80% of FTP and one where easy rides sit at 65% of FTP is larger than the difference between polarised and pyramidal. Execution within the model matters more than which model you pick.
5. Forgetting that recovery is part of the model. Every distribution includes rest. Polarised without rest days is just high-volume grey-zone training. Threshold-heavy without recovery weeks is a straight line to stagnation. The gap between sessions is where adaptation happens. Sleep is the most powerful recovery tool — not compression boots, not ice baths, not recovery shakes. Seiler's been clear about this, and the Ingebrigtsen camp puts sleep above every gadget in their hierarchy.
When to Switch Models
Three signals that your current distribution has run its course:
FTP plateau lasting 6+ weeks despite consistent training. The stimulus has been absorbed. Time to change the distribution — if you've been polarised, introduce a threshold-heavy build block. If you've been threshold-heavy, pull back to polarised and rebuild the aerobic base.
Chronic fatigue without performance gain. Heart rate climbing at the same power, RPE rising at threshold, motivation dropping. Usually means too much moderate intensity and not enough genuine recovery. Pull back to a strict polarised structure with one fewer intensity day.
Event demands have changed. Training for a time trial after a year of sportive preparation means the distribution needs to shift. The model follows the goal, not the other way round.
The Bottom Line
The polarised vs pyramidal debate is real, and the research has meaningful things to say. But the practical answer for most amateur cyclists isn't to pick a side — it's to use both at the right time.
Build your aerobic base with a polarised distribution. Develop race-specific fitness with a pyramidal build. Sharpen with targeted threshold and VO2max work. Recover properly between phases. Repeat, season after season, adjusting the percentages based on your available hours and your body's response.
The model that fails every time is the one most riders accidentally follow: moderate intensity, every ride, every week, no periodisation, no genuine easy days, no genuine hard days. Fix that and you'll outperform 90% of amateur cyclists regardless of whether you call your programme polarised, pyramidal, or anything else.
And if you want help applying this — if you'd rather have the distribution, the sessions, and the accountability built around your actual life — that's what the Roadman community does. Serious riders, honest training, no grey zone. Come join us on Skool. You're not done yet.