Polarized, Pyramidal, and Threshold Training: The Distribution Argument

Three intensity distributions, three coaching tribes. The polarised model has dominated endurance-training conversations since Seiler's 2010 paper; the pyramidal evidence from Casado has gradually rebalanced the debate; threshold training keeps losing ground but won't go away. This article unpacks where each model comes from, what the data actually shows, and where each one fits.

The formula and the framing

All three models partition total weekly training time into three intensity zones, anchored on the two lactate-threshold landmarks:

Zone 1: easy / aerobic         (below LT1, ~85% max HR or lower)
Zone 2: threshold / tempo      (between LT1 and LT2)
Zone 3: hard / VO2-max range   (above LT2)

The three distributions specify what fraction of weekly time falls into each zone:

Model        Zone 1   Zone 2   Zone 3
─────────────────────────────────────────
Polarised     80%      ~5%      ~15%
Pyramidal     75%      20%      5%
Threshold     50%      40%     10%

Derivation: how each model emerged

Polarised (Seiler 2010)

Seiler's polarised model was derived from elite cross-country skiers in the Norwegian national team. The cohort consistently spent ~80% of training time at very low intensity (sub-LT1) and ~15–20% at very high intensity (above LT2), with almost no time spent in the middle threshold zone. Seiler's argument was that the threshold zone produces high subjective effort without proportional adaptation — the "no man's land" that polarised training explicitly excludes.^[1]

Pyramidal (Casado 2019)

Casado and colleagues analysed training logs from elite marathon runners and found a different distribution: ~75% Zone 1, ~20% Zone 2, ~5% Zone 3. The middle-zone work was substantial and intentional, not accidental. Their argument: at marathon-distance volumes (often 150–250 km/week), the polarised 20% high-intensity allocation produces too much absolute time at unsustainable intensities. Pyramidal redistributes some of that hard work into the threshold zone, which the runners could sustain across the week.^[3]

Threshold (older default)

Threshold training dominated 1970s-90s endurance coaching, particularly for middle-distance and time-trial events. The intuition: train at the intensity you race at. The Stöggl & Sperlich 2014 trial pitted threshold against polarised and pyramidal under controlled total training stress; threshold under-performed both alternatives on race-pace improvement.^[2]

The empirical record

Three studies anchor the modern conversation:

1. Seiler 2010 — observational analysis of elite cross-country skiers. Established that high-performing endurance athletes naturally trend toward polarised distributions when given autonomy.^[1]
2. Stöggl & Sperlich 2014 — controlled 9-week intervention with three groups (polarised, threshold, high-volume). Polarised group improved VO2max significantly more than threshold at matched training load.^[2]
3. Casado et al. 2019 — retrospective analysis of top-tier marathoners. Found pyramidal distributions dominated above 100 km/week, regardless of nationality or coaching style.^[3]

The Esteve-Lanao et al. 2007 work was the early shot in the modern debate, comparing polarised vs threshold in sub-elite runners. Their results favoured polarised but at modest training volumes (~40 km/week); the boundary where pyramidal takes over was not yet clear in that data.^[4]

Where each model fails

Polarised at high volume

At 150 km/week running, the polarised 20% high-intensity allocation puts 30 km/week at race-pace or above. Even elite runners struggle to sustain that distribution across multi-week blocks. Above ~100 km/week, the polarised math forces the athlete into one of two compensations: drop total volume (defeating the high-volume strategy) or reclassify some "hard" work as threshold (sliding toward pyramidal).^[3]

Pyramidal in sub-elite populations

Pyramidal's 20% threshold allocation requires the athlete to have a stable, measured threshold pace. Sub-elite athletes whose threshold is poorly characterised will hit the wrong intensity in the threshold sessions, producing the no-man's-land effect Seiler warned against. Polarised is the safer default for athletes without lactate or breath-by-breath data.

Threshold under-performs in trials

Stöggl & Sperlich's controlled trial is the strongest evidence against pure threshold training. The threshold group accumulated similar total training stress as the polarised group but produced smaller VO2max and race-pace improvements. The mechanism is the published "threshold compromises both ends" effect: too hard for true aerobic adaptation, too easy to provoke maximal VO2 adaptation.^[2]

Bottom line: which model for which athlete

1. Recreational and sub-elite (under ~80 km/week): Polarised. The 80/20 ratio is achievable, the threshold zone is small enough to skip without losing benefit, and the model tolerates imprecise zone boundaries.
2. Trained athletes at moderate volume (80–120 km/week): Either polarised or pyramidal works. Personal preference, race-distance, and weekly schedule decide.
3. High-volume athletes (above ~120 km/week): Pyramidal. The 20% threshold allocation is necessary fuel placement at this volume.
4. Pure threshold: Avoid as a primary distribution. May appear during specific race-pace blocks (2–3 weeks pre-race) but should not anchor the year.

The practical implementation

All three models require accurate zone boundaries. The two anchors that matter most are LT1 (top of Zone 1) and LT2 (bottom of Zone 3). LT1 is approximately the highest pace that allows nasal-only breathing or full-sentence conversation; LT2 is the highest pace sustainable for ~30–60 minutes (functional threshold pace).

The Zone 2 Heart Rate Calculator handles the LT1-anchored Zone 1 ceiling. The Heart Rate Zone Calculator provides the full Karvonen zones for prescribing Zone 2 and Zone 3 work. The VO2 Max Estimator validates that the prescribed zones map onto actual physiological intensity.

Where the methodology bends

Time-in-zone vs session-in-zone

The polarised 80/20 ratio refers to time spent in each zone, not session count. A single 90-minute Zone 3 interval session contains 60 minutes of Zone 1 warm-up/cool-down — the in-zone time is the 30 minutes of intervals, not the full session duration. Athletes who count by session rather than by time consistently over-classify their distribution as more polarised than it actually is.^[1]

Cross-training effects

All three models were developed for single-sport endurance athletes. Cross-training (e.g., cycling sessions for a runner) complicates the distribution math because the heart-rate-vs-pace relationship shifts between disciplines. The published recommendation is to anchor each discipline's distribution to its own threshold landmarks rather than blending across modalities.^[4]

Block periodisation effects

The 80/20 (or 75/20/5) ratios describe annual averages, not weekly snapshots. During race-prep blocks, the distribution shifts toward more Zone 3 work for 3–6 weeks before reverting to base distribution during recovery. The annual average across blocks still hits the 80/20 target.

Worked weekly distribution by athlete profile

Three concrete weekly plans built off the three models, for an athlete with 6 training hours per week:

POLARISED (80/20)               PYRAMIDAL (75/20/5)            THRESHOLD (50/40/10)
─────────────────────────────────────────────────────────────────────────────────
Mon  60 min Z1 easy             Mon  60 min Z1 easy            Mon  45 min Z1 easy
Tue  45 min Z3 intervals        Tue  45 min Z2 tempo           Tue  60 min Z2 tempo
Wed  60 min Z1 easy             Wed  60 min Z1 easy            Wed  30 min Z1 easy
Thu  rest                       Thu  rest                      Thu  60 min Z2 tempo
Fri  30 min Z3 intervals        Fri  30 min Z3 intervals       Fri  rest
Sat  90 min Z1 long             Sat  90 min Z1 long            Sat  60 min Z2 tempo
Sun  75 min Z1 easy             Sun  75 min Z1 easy            Sun  90 min Z1 long
Total Z1 = 285 min (79%)        Total Z1 = 285 min (75%)       Total Z1 = 180 min (50%)
Total Z3 =  75 min (21%)        Total Z2 =  75 min (20%)       Total Z2 = 180 min (40%)
                                Total Z3 =  30 min (5%)        Total Z3 =  45 min (10%)

The threshold plan accumulates the most "moderate effort" hours per week but the published outcomes from Stöggl & Sperlich 2014 say this version produces the least adaptation per hour of any of the three.^[2] Most recreational athletes naturally drift toward the threshold plan because the threshold zone feels productive; the polarised plan feels easy on most days and uncomfortably hard on a few. The polarised plan's discomfort distribution is the active ingredient.

Cross-checking against related tools

The Zone 2 Heart Rate Calculator sets the Zone 1 ceiling. The Heart Rate Zone Calculator exposes all five Karvonen zones for prescribing the high-intensity slice. The VO2 Max Estimator validates that the assumed thresholds match physiological reality.

Related reading: Polarised vs Threshold Training: 2026 Systematic Review for the meta-analytic update, Zone 2 Training: What The Literature Says for the LT1-anchored aerobic-base framing, and Zone 2 Training: Methods and Limits for the methodological deep-dive on threshold measurement.

FAQ

Is 80/20 really the right ratio for me?

For recreational and sub-elite athletes at moderate volume, yes. The Seiler-derived 80/20 ratio is durable across a wide range of populations. For high-volume athletes (above 120 km/week running), pyramidal's 75/20/5 fits better because it places necessary threshold work without forcing it into an unsustainable 20% Zone 3 slot.^[3]

What happens if I do too much threshold?

The published trial outcome from Stöggl & Sperlich 2014: VO2max improves less than under polarised at matched training stress. The mechanism is sub-optimal recovery (threshold work blunts recovery from itself the way Zone 1 doesn't) without the maximal stimulus that Zone 3 provides.^[2]

How precise do the threshold boundaries need to be?

Polarised is forgiving (the middle zone is small, so misclassifying within ±5 bpm rarely matters). Pyramidal requires more precision because the 20% threshold allocation needs to land at actual threshold pace, not near it. For pyramidal athletes, a lactate test or breath-by-breath VO2 measurement every 12–16 weeks is well worth the cost.^[4]

Should I follow these models year-round?

Annual averages, yes. Within blocks, the distribution shifts. A typical 16-week marathon block might run 85/10/5 in weeks 1–8, then 70/20/10 in weeks 9–14, then 80/15/5 during taper. The annual average still hits the model's target ratio.

References

1. 1 Intensity distribution and performance in cross-country skiing (Seiler) — International Journal of Sports Physiology and Performance (2010)
2. 2 Polarized vs threshold training in elite endurance athletes (Stöggl & Sperlich) — Frontiers in Physiology (2014)
3. 3 Pyramidal training intensity distribution in top-tier marathon runners (Casado et al.) — Journal of Sport and Health Science (2019)
4. 4 Endurance training intensity distribution: review of three models (Esteve-Lanao et al.) — Journal of Strength and Conditioning Research (2007)