Polarised training is a training-intensity distribution: most endurance work sits below the first physiological threshold, a smaller amount sits above the second, and less work is accumulated in the moderate domain between them.
The popular shorthand is “80% easy, 20% hard.” That is a useful picture, but it is not a universal weekly prescription and it is not meaningful until three things are named:
- the zone model;
- the thresholds or anchors used to set it; and
- whether sessions or recorded time are being counted.
Without those details, two cyclists can both claim an 80/20 week while describing different training.
Polarised training at a glance
| Question | Useful answer | Common error |
|---|---|---|
| What is polarised training? | A distribution with most work low, some high and less moderate | Treating it as one workout or interval plan |
| Which zones? | A three-zone model separated by two physiological thresholds | Reading the numbers as seven-zone power labels |
| Is it exactly 80/20? | No; the ratio describes a broad pattern | Forcing every seven-day week to match exactly |
| Time or sessions? | Both are used, but they answer different questions | Mixing the methods in one comparison |
| Is moderate work useless? | No; value and cost depend on context | Turning “less moderate” into “never moderate” |
| Does it always win? | No; outcomes vary by population and programme | Promising universal superiority |
What polarised training means in the three-zone model
Polarised training research usually uses three intensity domains:
- Zone 1: low intensity, below the first ventilatory or lactate threshold;
- Zone 2: moderate intensity, between the first and second thresholds; and
- Zone 3: high intensity, above the second threshold.
These are not the same labels as a common seven-level cycling power system. Research Zone 1 contains the easy endurance work often called power Zone 2. Research Zone 3 can contain work that cycling software labels power Zone 4, Zone 5 or higher.
A review of training-intensity models shows why translation matters: three-zone, five-zone and other frameworks use different anchors and terminology. A zone number is incomplete without its model.
Use the FTP zone calculator for conventional power bands, but do not assume a fixed FTP percentage precisely locates the first or second physiological threshold. The cycling Zone 5 guide explains the same model problem at the high-intensity end.
Where 80/20 came from
The early work was descriptive: researchers examined how trained endurance athletes already organised their training.
In a 2006 study of well-trained junior cross-country skiers, Seiler and Kjerland compared several quantification methods. Heart-rate analysis produced about 75% in Zone 1, 8% in Zone 2 and 17% in Zone 3; session RPE produced a similar 76%, 6% and 18%. The sample was eleven male skiers over a limited observation period—not every cyclist in every phase.
Seiler's 2010 review described nationally and internationally competitive endurance athletes training ten to thirteen times per week. About 80% of sessions were low intensity and about 20% were dominated by high-intensity work. The conclusion emphasised high-volume low-intensity training with carefully applied high intensity.
That evidence supports a predominantly low-intensity shape. It does not prove that 80.0% is a biological optimum or that a four-ride recreational cyclist should manufacture one hard day to complete the ratio.
Is 80/20 measured by time or sessions?
Both approaches appear in research and practice:
- Session goal assigns the whole workout to the intensity that defines its main purpose. An interval ride can count as a high-intensity session even though its warm-up, recoveries and cool-down are low intensity.
- Time in zone counts recorded minutes in each domain, often using heart rate, power or another measured signal. The same interval ride can contain substantial low-intensity time.
The results are not interchangeable. Time-in-zone can also change with the signal: power responds immediately, heart rate lags, and thresholds derived from different tests can move the boundaries.
State the method beside the result. “Four low sessions and one high session” is a session-goal description. “Eighty per cent of heart-rate time below the first threshold” is a time-in-zone description. Neither should be presented as though it proves the other.
For a practical audit, use one method across the comparison period and record both when possible. The disagreement is useful information, not an error to hide.
Do elite cyclists really train polarised?
Elite endurance athletes perform most training at low intensity. The precise distribution is more varied than the slogan suggests.
A 2023 review of 175 elite-to-world-class training-intensity distributions found 89 pyramidal, 65 polarised and eight threshold distributions. The shape varied by sport and phase. Cyclists and swimmers showed less than 72% low-intensity work and more than 16% moderate work in that review, while several other endurance sports exceeded 80% low intensity and stayed below 12% moderate.
This does not mean professional cyclists ignore low-intensity volume. It means that “all elites are polarised” is inaccurate. Race demands, measurement method and seasonal phase affect the observed shape.
Roadman's WorldTour prescription article owns the narrower question of how elite coaching practice translates into a professional calendar. The broad definition and evidence remain here.
Does polarised training work better?
Polarised training can improve endurance performance. Whether it is superior depends on what it is compared with, which outcome is measured and who is training.
An earlier 2019 systematic review and meta-analysis found a moderate pooled time-trial effect favouring polarised over threshold-heavy training. Only four studies met the review criteria, three entered the meta-analysis and the included studies scored four or five out of ten on the PEDro quality scale.
A 2024 meta-analysis included seventeen studies and found a small average advantage for VO2peak, especially in shorter interventions and highly trained athletes. It found similar effects for time-trial performance, time to exhaustion and power or speed at the second threshold.
A cyclist-specific 2025 systematic review and meta-analysis included 41 studies, 81 training groups and 797 trained cyclists. Training improved VO2max and time-trial performance, but there was no significant difference between polarised and non-polarised categories.
An individual-participant network meta-analysis found no statistically significant overall difference between polarised and pyramidal interventions for VO2max or time-trial performance. Its subgroup result suggested competitive athletes may respond differently from recreational athletes.
Likewise, a 2023 review of periodisation and training-intensity distribution in cyclists found only seven eligible studies and no evidence that one model was superior across the reviewed eight-to-twelve-week interventions.
The evidence does not reduce to “polarised wins” or “distribution does not matter.” It says several distributions can work and that athlete level, comparison model, duration and outcome matter.
Is the grey zone useless?
No. Moderate training is not an empty physiological space.
The concern behind “grey zone” coaching is opportunity cost. A rider who unintentionally turns every low-intensity day moderate may accumulate more fatigue, reduce volume or arrive at demanding sessions unable to execute the intended work. That is a programming problem, not proof that moderate intensity produces no adaptation.
Pyramidal training deliberately includes more work in the moderate domain while keeping low intensity as the largest share. The elite review found pyramidal distributions more often than polarised ones overall. Event-specific tempo, sustained climbing and controlled sub-threshold work can all be legitimate uses.
Ask four questions:
- Was the moderate work planned or accidental?
- Does it match the event or current limiter?
- What low-intensity volume or high-intensity quality does it displace?
- Can the rider absorb it and repeat the rest of the programme?
The sweet spot training guide applies those boundaries to 88–94% FTP. The polarised vs sweet spot comparison handles the narrower choice between a distribution and a session type.
How to implement polarised training for cycling
Start with an audit, not a ratio.
1. Define the model and thresholds
Use a three-zone framework and record how the first and second boundaries were estimated. Laboratory lactate or ventilatory testing can provide individual anchors; field estimates, talk test, RPE, heart rate and power can support practical decisions but have different limitations.
Do not call conventional power Zone 2 “research Zone 2”. Translate the labels explicitly.
2. Choose the counting method
Use session goal, time in zone or both. Keep the method stable for several representative weeks. A single recovery week, race week or travel week is not the athlete's normal distribution.
3. Include the full load
Count races, hard group rides, commuting surges, running, demanding strength work and other endurance sports. A Sunday race is high-intensity training even when no interval workout appears in the plan.
4. Protect genuinely low-intensity work
Low-intensity riding should serve its purpose without becoming an FTP-percentage contest. Use the talk test, breathing, RPE and an individual heart-rate or power ceiling as supporting signals. Heat, altitude, fatigue and medication can change the internal cost at the same watts.
5. Add high intensity for a reason
Select the hard workout from the goal. VO2max formats, threshold work, races and short high-intensity intervals are not interchangeable just because all fall in research Zone 3. Use the VO2max interval guide or threshold interval guide for the specific job.
There is no universal requirement for one hard session per four easy sessions or two hard days every week. Add the smallest dose that can be executed with repeatable quality.
6. Review outcomes, not ratio compliance
Track comparable session output, event-specific performance, low-intensity durability and recovery. A perfect pie chart with declining performance is not success. A slightly pyramidal distribution with improving, repeatable training is not failure.
Example weeks without false precision
These are classification examples, not prescriptions.
Four-ride week
- one demanding interval session with a defined goal;
- two low-intensity endurance rides;
- one longer ride whose intensity follows the event phase.
This may look polarised by session goal while appearing more pyramidal by recorded time if the long ride includes planned tempo. Do not add another hard day solely to produce 20%.
Six-ride week with a race
- one race, already counted as high intensity;
- four low-intensity or recovery-purpose rides;
- one event-specific session that may be low, moderate or high depending on the phase.
The race changes the week. A template that ignores it will over-prescribe intensity.
Multi-sport or strength week
- count hard running and demanding lower-body strength before adding cycling intensity;
- classify the cycling sessions separately and the complete athlete load together;
- use the combined recovery response to decide the next demanding session.
Polarised cycling data can look tidy while the athlete's total programme is not polarised at all.
Polarised vs pyramidal training
Both distributions contain mostly low-intensity work. The distinction is the ordering of the smaller domains:
- polarised: low intensity is largest, high intensity is next, moderate is smallest;
- pyramidal: low intensity is largest, moderate is next, high intensity is smallest.
The polarised vs pyramidal guide owns that comparison and already has a distinct Search Console footprint. Do not collapse comparison intent into the broad definition page.
Common mistakes
- Mixing zone systems. Research Zone 1 is not power Zone 1, and research Zone 3 is not cycling power Zone 3.
- Mixing counting methods. A session-goal ratio cannot be compared directly with heart-rate time in zone.
- Forcing the ratio weekly. Training distribution is more interpretable across representative blocks than one disrupted seven-day period.
- Calling all moderate work junk. The problem may be accidental dose or opportunity cost, not the intensity's existence.
- Ignoring races and group rides. They can already supply the high-intensity share.
- Promising universal superiority. Current reviews do not support one winner for every cyclist and outcome.
- Using weekly hours as the only selector. Performance level, event demands, phase and previous response also matter.
When to change the distribution
Change it when evidence from the rider shows a problem:
- low-intensity volume repeatedly becomes unintended moderate work;
- demanding sessions lose repeatable quality;
- event-specific performance is not being trained;
- recovery trends remain abnormal;
- the distribution no longer matches the current phase; or
- the rider is healthy but has stopped responding to a previously effective pattern.
Change one variable at a time. Moving from threshold-heavy to polarised training while also increasing volume and changing every interval prevents you from knowing what caused the response.
Persistent chest pain, fainting, unusual breathlessness, palpitations or other concerning exercise symptoms require appropriate medical assessment, not a different intensity distribution.
The Roadman decision rule
Use polarised training as a way to organise contrast, not as an identity or a weekly score.
Define the three-zone model. State whether you count sessions or time. Audit the actual programme. Protect low-intensity work when it serves volume and recovery. Add high intensity only for a clear purpose. Allow moderate work when it is deliberate and useful.
Then review performance and recovery rather than asking whether the chart is exactly 80/20.
That is the trusted version of polarised training: predominantly low intensity, carefully applied high intensity, less unplanned moderate work—and no claim that one ratio wins for every cyclist.