Polarized training: definition, formula, example

Polarized training describes a way of spreading training intensity: most of the volume clearly easy, a smaller but real share clearly hard, and very little in the middle. The idea did not come from a lab protocol but from watching what elite endurance athletes actually do. Reviewing the descriptive studies, Stephen Seiler found that nationally and internationally competitive endurance athletes training 10 to 13 times a week tend to settle on about 80% of their sessions at low intensity and about 20% dominated by hard work, such as intervals at around 90% of VO2max (Seiler, 2010).

That pattern is often shortened to “80/20”. The short version hides two details that matter: which zones you count, and whether you count sessions or minutes.

The three zones behind it

Polarized training uses a three-zone model built on two physiological thresholds. Zone 1 sits below the first ventilatory or lactate threshold (blood lactate around 2 mmol/L): easy running where you can talk in full sentences. Zone 2 lies between the two thresholds. Zone 3 is above the second threshold (around 4 mmol/L, close to critical speed), where intervals live.

Watch out for the name clash. In the five-zone systems on most watches, “zone 2” means easy endurance, which falls inside zone 1 of the three-zone model. Polarized training does not tell you to avoid easy zone 2 runs; it tells you to spend little time between the thresholds.

Putting a number on it

Treff and colleagues (2019) proposed a polarization index to tell a polarized distribution from the others, using the fraction of training time (or distance) in each of the three zones:

A worked example

A runner trains 8 hours a week. Over a month, time in zones averages 80% in zone 1, 5% in zone 2 and 15% in zone 3.

• Z1 / Z2 = 0.80 / 0.05 = 16
• 16 × 0.15 × 100 = 240
• PI = log10(240) = 2.38, above 2.00: polarized

Change the mix to 75% / 20% / 5%, a pyramid with plenty of tempo running and few intervals: PI = log10(0.75 / 0.20 × 0.05 × 100) = log10(18.75) = 1.27. A threshold-heavy block at 60% / 35% / 5% gives 0.93. Same weekly hours, three very different weeks.

Sessions or minutes

Seiler’s 80/20 counts sessions. Time in zone gives a different picture, because a hard session is mostly warm-up, recoveries and cool-down. A 70-minute session of 5 × 4' at 3K to 5K pace has 20 minutes of work, and with heart rate lagging behind pace, fewer than that are recorded in zone 3. Seiler and Kjerland (2006) measured both ways in eleven junior cross-country skiers over 32 days: by heart rate, their endurance sessions split 75% / 8% / 17% across the three zones, and session RPE gave almost the same result. Four easy hour-long runs plus that interval session make a week that is 80/20 by sessions, but only 20 of its 310 minutes, about 6%, are hard. Neither is wrong, but compare like with like.

How to use it

For a runner with five or six sessions a week, a polarized week can look like this: three or four easy runs and a long run, all below the first threshold, plus one or two hard sessions such as 5 × 4' or 6 × 3' at 3K to 5K pace with equal jog recoveries. The easy days have to stay easy, which for most people means slower than they would like; set the ceiling with the heart rate zone calculator or the talk test. OpenAthlete’s statistics show the share of training below, between and above your thresholds, which is enough to check which kind of week you actually ran.

Limits

The strongest controlled support comes from Stöggl and Sperlich (2014): 48 well-trained runners, cyclists, triathletes and skiers spent nine weeks on one of four programmes, and the polarized group improved VO2peak by 11.7% and time to exhaustion by 17.4%, more than the high-volume, threshold or interval groups. Nine weeks and 48 athletes split across four groups is a short and small study, and its subjects were already fit. Seiler (2010) himself frames the 80/20 pattern as descriptive: it is what successful athletes converge on, not proof that it is optimal for everyone. A runner with four hours a week may gain more from some threshold work than from strict polarization; that question is still open. Finally, the zones depend on thresholds you have to test or estimate, and a 10 bpm error moves minutes from one zone to the next.