Polarised, Pyramidal or Threshold: How to Spread Your Winter Intensity

If you've read anything about endurance training in the last decade, you've probably been told to train polarised: around 80% easy, 20% hard, and as little as possible in the "grey zone" in between. It's become one of the most repeated ideas in endurance sport. It's also more contested than the slogans suggest.

This article looks at what intensity distribution means, what the trials comparing distributions have found, and what the newest and largest analyses say. It focuses on the whole week rather than individual sessions. TriForward's "Raising FTP on 8 Hours a Week" covers which intervals to choose, and "Double Threshold, Single Reality" covers the Norwegian method.

What intensity distribution means

Most research uses a three-zone model, anchored to two physiological thresholds:

• Zone 1: below the first ventilatory or lactate threshold. Easy, conversational riding or running.
• Zone 2: between the two thresholds. Steady, "tempo" or "sweet spot" work. Note that this is not the same as the popular five-zone "Zone 2", which sits within the research Zone 1. TriForward's "Finding Zone 2: Heart Rate, Lactate, Talk Test or Breathing Threshold?" covers that confusion.
• Zone 3: above the second threshold. Hard intervals.

Intensity distribution is how your training time splits across these zones. The common patterns are:

• Polarised: most time in Zone 1, a meaningful share in Zone 3, and little in Zone 2.
• Pyramidal: most time in Zone 1, less in Zone 2, and least in Zone 3.
• Threshold: a large share of time in Zone 2.
• High-intensity: a large share in Zone 3.

The labels sound precise, but studies define them slightly differently, so two "polarised" programmes can look quite different in practice.

The trials that made polarised training famous

Mixed endurance athletes

A 2014 study by Stöggl and Sperlich in Frontiers in Physiology randomised 48 well-trained endurance athletes (runners, cyclists, triathletes and cross-country skiers, aged 31 ± 6) to nine weeks of one of four distributions. The sex split wasn't reported.

• Polarised (68% low, 6% threshold, 26% high): VO2peak rose 11.7% and power or speed at a blood lactate of 4 mmol/L rose 8.1%.
• High-intensity (43% low, 57% high): VO2peak rose 4.8%.
• Threshold (46% low, 54% threshold): no significant change in VO2peak, and power or speed at 4 mmol/L rose only 1.4%.
• High-volume (83% low): no significant change in VO2peak.

The threshold group also had the most dropouts: four of the seven athletes who withdrew. These were small groups, and the athletes' sports were reported inconsistently within the paper, but the result became widely quoted.

Trained cyclists

A 2013 crossover study by Neal and colleagues in the Journal of Applied Physiology gave 12 trained male cyclists two six-week blocks, separated by a four-week break.

• Polarised block: about 6.4 hours a week, 80% low and 20% high intensity.
• Threshold block: about 7.5 hours a week, 57% low and 43% moderate intensity.

Both blocks improved performance, but polarised training improved more:

• 40 km time trial: improved 8% after the polarised block against 3% after the threshold block.
• Power at lactate threshold: improved 9% against 4%.
• Time to exhaustion at 95% of maximal work rate: improved 85% against 37%.

The polarised block also involved about an hour less training a week. That's an encouraging result for time-limited athletes, but it's 12 men over six weeks.

Recreational runners

The trial closest to age-group reality is a 2014 study by Muñoz, Seiler and colleagues in the International Journal of Sports Physiology and Performance. Thirty recreational runners (sex not stated in the abstract) trained for 10 weeks in one of two ways:

• Polarised: 77% in Zone 1, 3% in Zone 2, 20% in Zone 3.
• Between thresholds: 46% in Zone 1, 35% in Zone 2, 19% in Zone 3.

Both groups got faster over 10 km. The polarised group improved by 119 seconds (5.0%) and the between-thresholds group by 84 seconds (3.6%), but the difference between groups wasn't statistically significant. A subset analysis of runners who stuck most closely to their plans favoured polarised training, but that was a small, after-the-fact comparison.

What the bigger analyses show

Individual trials are small, so pooled analyses are more informative.

A 2024 meta-analysis

Oliveira, Boppre and Fonseca published a systematic review and meta-analysis in Sports Medicine in 2024. It included 17 studies (14 in the meta-analysis) with 437 participants: 317 men, 89 women and 31 not reported, from trained to highly trained or national level.

• VO2peak: polarised training produced a small advantage, a standardised mean difference (SMD) of 0.24.
• Time-trial performance: no difference (SMD −0.01).
• Threshold power or speed and time to exhaustion: no difference.
• Subgroups: the VO2peak advantage appeared in interventions shorter than 12 weeks (SMD 0.40) and in highly trained or national-level athletes (SMD 0.46). It disappeared in interventions of 12 weeks or longer and in trained or developmental athletes.

The authors noted that these two subgroups overlap, because highly trained athletes were mostly studied in short interventions, so it isn't possible to say which factor matters. They suggested polarised training may be useful for quick VO2peak gains before competition.

A 2025 analysis of individual athletes' data

The most detailed analysis so far is a 2025 network meta-analysis by Rosenblat and colleagues in Sports Medicine, with Stephen Seiler, who popularised the polarised concept, as senior author. Instead of pooling study averages, the authors collected individual participant data from 13 studies: 348 endurance athletes, 296 men and 52 women, of whom 150 were recreational and 198 competitive. The sports were cross-country skiing, cycling, rowing, running and triathlon.

• Polarised versus pyramidal: no significant difference in VO2peak (SMD −0.06) or time-trial performance (SMD −0.05).
• Polarised versus threshold, high-intensity or low-intensity: no significant differences either, though these groups were small.
• Training level: for VO2peak, competitive athletes trended towards polarised training and recreational athletes towards pyramidal training. The difference between those trends was significant, but neither trend on its own was.
• Study length and sport: made no significant difference.

The authors concluded that no distribution was clearly superior, and suggested that competitive athletes may benefit more from polarised training and recreational athletes from pyramidal training for VO2max. That's a hypothesis to test, not a finding.

Order may matter more than labels

A 2022 study by Filipas, Bonato, Gallo and Codella in the Scandinavian Journal of Medicine and Science in Sports added a twist. Sixty well-trained male runners followed one of four 16-week plans with the same training load: pyramidal throughout, polarised throughout, pyramidal then polarised, or polarised then pyramidal.

• All four plans improved endurance markers and performance.
• Pyramidal followed by polarised improved most: about 3.0% in relative VO2peak, 1.5% in lactate-threshold speed and 1.5% in 5 km time-trial performance.

Inference: a winter built on easy and steady work, followed by a shift towards more high-intensity work as the season approaches, fits this result. TriForward's "Block Periodisation vs Traditional Build: Does the Structure of Your Winter Matter?" looks at structure in more detail.

Why the early trials looked so different

Several reasons may explain why early trials found big advantages and pooled analyses don't:

• Small samples. With 12 to 48 athletes split into groups, a few strong responders can drive a large average difference.
• Different total training. In Neal's study, the threshold block had more hours. Some trials changed volume as well as distribution.
• Measurement. Most studies count time in heart-rate zones. Heart rate lags during short intervals, so a hard interval session can register mostly as Zone 1 or 2. The same week can look polarised or pyramidal depending on how it's counted.
• Adherence. Muñoz's subset analysis suggests that how closely athletes follow a plan matters.

Counting your own distribution

Before changing your distribution, it's worth knowing what it is. Many triathletes believe they train polarised and find, when they look, that most sessions drift into the middle. A simple check over a typical fortnight:

1. Label each session by its purpose before you do it: easy, steady or hard. A session with 5 x 4 minutes hard counts as a hard session, even though most of its minutes are warm-up, recoveries and cool-down.
2. Check the easy sessions afterwards. If heart rate or power on "easy" days sits well above your first threshold for long stretches, they weren't easy.
3. Count across all three sports. A hard run on Tuesday and hard bike intervals on Wednesday are two hard sessions for the same body.
4. Compare sessions with minutes. Time-in-zone from heart rate will usually make a week look more easy-dominated than counting sessions does.

Inference: if most of your sessions are labelled steady, your week is closer to the threshold pattern that did worst in the small trials than to either polarised or pyramidal training. That's the most useful thing to fix, whichever model you then choose.

What this means for an age-group triathlete

Most of the research is in single-sport athletes. A triathlete's week is different:

• Three sports share one recovery budget. A hard run and a hard bike session both draw on it.
• Swimming intensity can be hard to classify, because heart rate behaves differently in the water and most zone research is on land.
• Time is limited. On 8 to 10 hours a week, a strict polarised split leaves only a couple of hours of hard work, which is probably fine. A pyramidal split adds some steady work, which is also fine.
• Race demands differ. Inference: a half or full Ironman is a long, sub-threshold effort, so steady work is closer to race pace for long-course athletes than short, hard intervals are.

Inference: the evidence doesn't support strong rules about the exact percentages. It does support keeping most training easy and avoiding weeks dominated by threshold work.

How to spread your winter intensity

The following combines the trials above. It's inference for triathletes, not a tested plan:

1. Keep most of your training easy across all three sports. The polarised groups in these trials spent 68 to 80% of their time at low intensity.
2. Include one or two genuinely hard sessions a week, spread across the bike and run.
3. Use steady or threshold work deliberately, especially if you race long course, but don't let it become most of your week.
4. Consider shifting the emphasis over the winter. Based on Filipas, a pyramidal base followed by a more polarised block before the season is a reasonable pattern.
5. Count sessions, not just minutes. Decide each session's purpose in advance rather than relying on time-in-zone from heart rate.
6. Watch recovery. If hard sessions start to suffer, reduce intensity. TriForward's "Reading the Trend, Not the Number" covers monitoring.

A triathlon scenario

This example is illustrative, not drawn from a study. A 42-year-old training for a full-distance race next August has about nine hours a week over winter. Last winter, almost every bike session was a 60-minute "sweet spot" workout on the trainer, and most runs drifted into a steady effort. By February she was tired and her threshold hadn't moved.

A better winter, based on the evidence above:

• November to January: a pyramidal week. Easy swims, one steady bike session of 60 to 90 minutes, one shorter interval session on the bike, an easy long run and one run with some steady segments. Everything else easy.
• February to March: shift towards polarised. Keep the easy volume, swap the steady bike session for a harder interval session, and add short hard efforts to one run.
• From April: add longer race-pace work for the full distance.

Neither phase is right or wrong. The point is that most of her training becomes clearly easy, and the hard sessions become genuinely hard.

Common mistakes

• Making every session moderate. This is the threshold-heavy pattern that did worst in the small trials.
• Doing "easy" sessions too hard, which drags a polarised plan into the middle.
• Obsessing over exact percentages. The large analyses suggest the precise split matters less than consistency.
• Copying elite volumes. A professional's 80% easy is 20 hours; yours may be 6.
• Counting only bike sessions and ignoring hard runs and swims.

What we don't know

• Triathletes. Very few trials included triathletes training three sports at once.
• Women. Only 52 of the 348 athletes in the largest analysis, and 89 of 437 in the 2024 meta-analysis, were women.
• Long-term effects. Most trials lasted under 12 weeks, and none followed athletes across a full season.
• Long-course racing. Most performance tests were short time trials, not half or full Ironman efforts.
• The best measure of intensity. Different ways of counting zones can change how a week is classified.

References

• Rosenblat MA, Watt JA, Arnold JI, Treff G, Sandbakk ØB, Esteve-Lanao J, et al. Which training intensity distribution intervention will produce the greatest improvements in maximal oxygen uptake and time-trial performance in endurance athletes? A systematic review and network meta-analysis of individual participant data. Sports Medicine, 2025;55(3):655–673. doi:10.1007/s40279-024-02149-3
• Oliveira PS, Boppre G, Fonseca H. Comparison of polarized versus other types of endurance training intensity distribution on athletes' endurance performance: a systematic review with meta-analysis. Sports Medicine, 2024;54(8):2071–2095. doi:10.1007/s40279-024-02034-z
• Stöggl T, Sperlich B. Polarized training has greater impact on key endurance variables than threshold, high intensity, or high volume training. Frontiers in Physiology, 2014;5:33. doi:10.3389/fphys.2014.00033
• Neal CM, Hunter AM, Brennan L, O'Sullivan A, Hamilton DL, De Vito G, Galloway SDR. Six weeks of a polarized training-intensity distribution leads to greater physiological and performance adaptations than a threshold model in trained cyclists. Journal of Applied Physiology, 2013;114(4):461–471. doi:10.1152/japplphysiol.00652.2012
• Muñoz I, Seiler S, Bautista J, España J, Larumbe E, Esteve-Lanao J. Does polarized training improve performance in recreational runners? International Journal of Sports Physiology and Performance, 2014;9(2):265–272. doi:10.1123/ijspp.2012-0350
• Filipas L, Bonato M, Gallo G, Codella R. Effects of 16 weeks of pyramidal and polarized training intensity distributions in well-trained endurance runners. Scandinavian Journal of Medicine and Science in Sports, 2022;32(3):498–511. doi:10.1111/sms.14101