Every training forum on the internet hosts the same argument. One camp insists on polarised — 80% easy, 20% hard, nothing in the middle. The other defends sweet spot — stack 88–93% of FTP, accumulate the work, get fit fast. Both sides cite studies. Both sides call the other wrong.
The honest answer is that both methods work, neither is universally optimal, and the research is more specific than the internet makes it sound. Prof. Stephen Seiler's polarised work was done on elite endurance athletes. Dr. Andrew Coggan's sweet spot concept was built for time-constrained riders. They were solving different problems for different populations.
This article covers what each method actually does, what the studies actually concluded, and how to pick the one that matches your weekly hours, your training age, and your goal event.
The two training philosophies in one paragraph each
Polarised training distributes roughly 80% of sessions below the first lactate threshold (LT1, roughly 60–75% of FTP) and 20% above the second lactate threshold (LT2, above 90% of FTP). The middle zone — tempo and threshold — is deliberately minimised. The logic is that easy work builds aerobic base without generating fatigue, and hard work drives VO2 and neuromuscular adaptation. Moderate work sits in a no-man's-land: too hard to recover from, not hard enough to force adaptation. Prof. Stephen Seiler formalised this after observing elite Scandinavian endurance athletes across skiing, rowing, and running.
Sweet spot training typically targets approximately 88-93% of FTP (often described in ranges up to ~93% of threshold power — a tight band at the upper end of Zone 3/lower Zone 4 in the Coggan seven-zone model). Sessions run 2x20, 3x15, or 4x12 minutes with short recoveries. The zone was popularised by Dr. Andrew Coggan and Hunter Allen (Training and Racing with a Power Meter) as the intersection of high training stress and manageable fatigue. For a rider with 6–8 hours per week, sweet spot packs more metabolic adaptation into the available time than most other distributions. You can calibrate these zones using our FTP zones calculator once you have a recent test result.
"The 80/20 split is about session distribution, not time distribution. One hard interval session counts as one session, even if the intervals are only 12 minutes of a 90-minute ride. People reading '20% hard' as '20% of total riding time at high power' end up wildly over-cooked."
— Professor Stephen Seiler, exercise physiologist (Roadman Cycling Podcast)
What Prof. Seiler's polarised research actually found
Seiler's work started as observation. He tracked elite cross-country skiers, rowers, and runners and found a consistent pattern: approximately 80% of training sessions were performed at low intensity, 20% at high intensity, with very little time at threshold (Seiler & Kjerland, 2006; Seiler, 2010). This was the descriptive finding — what elites actually do, not what he prescribed.
The intervention studies came later. Stöggl and Sperlich (2014, published in Frontiers in Physiology) compared polarised, threshold, high-volume, and high-intensity groups over 9 weeks in 48 well-trained endurance athletes — runners, cyclists, triathletes and cross-country skiers. The polarised group produced the largest improvements in VO2peak, time to exhaustion, and peak power. This is the study most often cited when people say "polarised is better."
Two things get lost in the retelling. First, the athletes in these trials were already trained and already accumulating significant volume. Second, the comparison groups included threshold-heavy protocols that nobody with a decent coach would prescribe in isolation. The polarised advantage was real but contextual.
Seiler has said on the podcast and in published interviews that the 80/20 split describes session distribution, not time distribution. A 20-minute VO2 interval session counts as one hard session even though the intervals themselves might be 12 minutes of a 90-minute ride. This matters. People reading "20% hard" as "20% of your time at high power" end up wildly over-cooked.
The research supports polarised distributions for athletes training more than 10–12 hours per week. Below that volume, the 80% easy becomes too compressed to drive meaningful aerobic adaptation, and the method loses its edge.
What sweet spot training actually does
Sweet spot produces three adaptations efficiently: increased lactate clearance, raised functional threshold power, and improved muscular endurance at near-race intensities. Coggan's original framing was pragmatic — if a rider only has 45–90 minutes, sweet spot intervals deliver more training stress score per hour than either easy riding or VO2 work.
At 88–93% of FTP, you're producing lactate faster than at tempo but clearing it fast enough to sustain 20–40 minutes of total work per session. Mitochondrial density, capillarisation, and lactate transporter expression all respond. The fatigue cost is lower than threshold or VO2 work, which means you can do sweet spot two or three times a week without blowing up.
The limitation is ceiling. Sweet spot raises FTP effectively for the first 12–18 months of structured training, then the returns diminish sharply. Riders who plateau on sweet spot usually need either more volume (which forces a polarised distribution by default) or more time above FTP to push VO2max higher.
Sweet spot also has a known failure mode: it drifts. A session prescribed at 90% of FTP becomes 85% when you're tired, then 82%, then you're just riding tempo for an hour. Tempo riding is the most overrated training stimulus in cycling. It's hard enough to generate fatigue, not hard enough to force adaptation. If your sweet spot sessions consistently drift below 85%, you're not doing sweet spot — you're doing the middle zone Seiler specifically warns against.
This is one of the most common patterns we see in riders coming to our coaching from self-directed plans. They think they're doing sweet spot. The power files say otherwise.
The time-crunched cyclist problem
Here's where the polarised-vs-sweet-spot debate gets distorted. Most amateur cyclists train 6–10 hours per week. Polarised distributions were validated at 15–25 hours per week. The maths doesn't transfer cleanly.
If you train 20 hours a week, 80% easy equals 16 hours of base building. That's enough low-intensity volume to drive significant mitochondrial and cardiovascular adaptation on its own. The remaining 20% — roughly 4 hours of hard work per week — sharpens the engine.
If you train 7 hours a week, 80% easy equals 5.6 hours. That's not enough low-intensity work to move the needle for most trained cyclists. The 20% hard portion — 1.4 hours — isn't enough either. You end up doing too little of everything.
Dan Lorang's WorldTour athletes at Red Bull–Bora–Hansgrohe train high volumes in peak weeks — often in the 25-plus hour range. Their polarised distribution makes physiological sense because the absolute volume of easy riding is substantial. For a rider doing 8 hours, a modified approach — something closer to 70% low / 10% tempo-sweet spot / 20% high — often produces better results than strict 80/20.
Joe Friel has written about this for two decades. The principle is that intensity distribution should scale with volume. Low volume demands higher relative intensity. High volume permits more easy riding because the absolute dose is still adequate. Below 10 hours per week, sweet spot blocks of 6–10 weeks usually produce faster FTP gains than strict polarised programming.
The exception is riders returning from injury, older athletes managing recovery, or anyone with high life stress. In those cases, polarised distributions reduce systemic load even at lower volumes because the hard days are hard and the easy days stop accumulating hidden fatigue.
Which method works for your situation
Use polarised when you train more than 10–12 hours per week, when you're targeting events longer than 3 hours, when you have 6+ months to goal event, or when recovery is your limiter. The method rewards volume and punishes compressed schedules. It also tends to protect long-term motivation because the easy days stay easy and the hard days stay meaningful.
Use sweet spot when you train 5–9 hours per week, when your goal event is under 2 hours, when you have 8–16 weeks to goal event, or when FTP is the clear limiter. It produces faster adaptations in the first year of structured training and fits around work schedules because sessions fit inside 90 minutes.
Training age matters more than most riders admit. A cyclist in their first three years of structured training responds to almost any reasonable protocol. Sweet spot often wins simply because it's efficient and the FTP gains are large. A cyclist with 5+ years of structured training has exhausted the easy gains from threshold-adjacent work. Polarised distributions and true VO2 work become the productive stimulus.
Event specificity should override method loyalty. A 40k TT demands different preparation than a 160k gran fondo with 3,000m of climbing. Dan Bigham — former UCI Hour Record holder, now Head of Engineering at Red Bull–Bora–Hansgrohe — has described on the podcast how track pursuit preparation looks nothing like a grand tour build. The intensity distribution is downstream of the demand.
If you're coaching yourself, the honest test is your power file from the last 30 days. Open the distribution. If 40% of your time sits in zones 3–4 (tempo and threshold) and your FTP has been static for six months, you're doing the worst of both worlds. Either commit to polarised and push more time into zone 2, or commit to sweet spot and make sure those intervals actually hit 88–93%.
How to combine both without wrecking yourself
The competent periodised approach uses both methods at different phases. Base phase (12–16 weeks out) runs polarised — high zone 2 volume, one VO2 session per week to maintain top-end. Build phase (6–10 weeks out) introduces sweet spot blocks to raise sustained power. Specificity phase (3–6 weeks out) narrows to race-specific intensities.
The non-negotiable rule is not stacking threshold and VO2 work in the same microcycle. Pick one intensity driver per week. If Tuesday is VO2 intervals, Thursday is sweet spot or tempo at most, not another threshold smash. If Tuesday is 3x15 at sweet spot, Thursday can be VO2 because the residual fatigue is lower.
Easy days must be easy. John Wakefield, Director of Coaching & Sports Science at Red Bull–Bora–Hansgrohe and founder of Science to Sport, has talked about this on the podcast — the most common error in amateur programming is creeping intensity on recovery days. A zone 2 ride at the top of zone 2 is still a zone 2 ride in theory and a half-tempo ride in practice. Cap recovery rides at the lower end of zone 2, not at the top.
Track two metrics across an 8-week block: weekly time in zone 2 (should be rising or stable) and peak 20-minute power (should be rising or stable). If both are falling, you're over-reaching. If zone 2 time is falling but peak power is rising, you're drifting into a threshold-heavy pattern that will stall within 4–6 weeks.
For triathletes, the bike intensity distribution has to protect the run. High-threshold bike volume wrecks run quality within 48 hours. Polarised bike work preserves running legs better than sweet spot, which is why most triathlon bike coaching uses polarised distributions even when pure cyclists in the same training volume would benefit from sweet spot.
Pick one method for the next 6 weeks. Log every session with power. At the end of the block, check your FTP, your zone 2 pace at fixed heart rate, and your peak 5-minute power. The numbers will tell you what's working.
What the more recent research adds
Filipas et al. (2022), Scandinavian Journal of Medicine & Science in Sports. Sixty well-trained male runners were assigned to four 16-week interventions: pyramidal (PYR), polarized (POL), PYR→POL, or POL→PYR. The PYR→POL sequence produced the largest improvements in VO2peak, lactate-threshold velocities and 5 km time-trial performance. Pyramidal distributions (which include sweet-spot intensities) work in base, but switching to polarized in the build phase amplified gains.
Magalhães et al. (2024), Sports (Basel). Sixteen weeks of pyramidal-distribution training in recreational male cyclists produced significant gains in power at fixed lactate concentrations. Time accumulated in zone 2 was the strongest correlate of performance change — a useful sanity-check that the time-crunched amateur population responds well to pyramidal/sweet-spot-inclusive distributions, not only polarized ones.
Wong, Burnley et al. (2022), Journal of Sports Sciences. FTP is not a valid marker of the maximal metabolic steady state. In this study, time to exhaustion at FTP averaged 33.7 minutes (not 60), and FTP+15W produced clear physiological drift. The implication: sweet spot zones calculated from FTP may land below or above the true sustainable intensity for an individual rider. If your sweet spot sessions feel too easy or too hard, the FTP test itself may be miscalibrated.
For the complete research library, see the research & evidence hub. For the source-level reads, see what Stephen Seiler says about polarised training, Stephen Seiler research lessons, the polarised training cycling guide, the sweet spot training cycling guide, and the zone 2 complete guide.
Got a specific question about which model fits your situation — your event, your hours, your history? Ask Roadman for an answer drawn from the actual coach and physiologist conversations on the podcast. And if you'd rather have someone build the right model around your week, the application for NDY coaching is where that starts.
