SYNTH-26 · Guide

Intensity distribution is a function of HOURS, not of ambition

What to do with 4, 8 and 15 hours a week

Blimited evidenceMostly from recreational to highly trained athletesWritten before October 2026, being re-checked

The decision this framework makes

Given an athlete's REAL weekly hours, how much of that time is easy, how much is hard, and how many hard sessions the week can carry. Every later choice — interval format, rest placement, ramp — sits on top of this one.

Why the standard answer is wrong for most athletes

The polarized/pyramidal evidence base is overwhelmingly elite and high-volume: young world-class rowers at ~95% of rowing below 2 mmol/L lactate (REF-638), Olympic speed skaters across 38 years shifting toward more zone-1 and less zone-2/3 time (REF-670), highly trained swimmers whose distribution follows the event — sprinters polarized/threshold, middle-distance threshold/pyramidal, distance pyramidal (REF-650, a qualitative review of 9 studies), and the fastest marathoners in a dataset of 119,452 runners, who ran more than three times the volume of slower runners, almost all of the extra in Z1, with >80% of them pyramidal (REF-197).These describe what 15–25 h/week looks like when it works (the "15–25 h" is our reading of these cohorts, not a number any of them reports).
80/20 applied to 5 hours means 4 easy hours and 1 hard hour. Four hours of easy riding is below the volume at which low-intensity work does its job, and one hard hour is one session — too thin to progress. The ratio is preserved and the training is gone. A distribution is not scale-invariant. (Inference: no study tests this arithmetic directly.)
What the dose-response evidence actually says about the low end: per hour of training, higher intensity buys more mitochondrial adaptation — HIT ~1.7× and SIT ~3.9× the gain of low-intensity work per hour — while total gains were similar across modalities after adjusting for frequency, weeks and initial fitness (pooled from untrained and diseased to trained people; muscle-biopsy markers, not performance), and more sessions per week (6 > 4 > 2) means more gain (REF-465, A, a mitochondrial/capillary meta-regression, 353 + 131 articles, 5,973 participants).That paper does NOT say how many hard sessions a week an athlete tolerates — it says the per-hour return favours intensity and the per-week return favours frequency. Sessions, not percentages, are therefore the right currency when hours are scarce.

The framework — by available hours

Under ~5 h/week. Percentages are meaningless here; count sessions. 2 quality sessions + whatever easy riding fits. Not from a study; the base's own practice: the "2" is a recovery-budget judgement, not a measured optimum.

The quality is what buys mitochondrial adaptation per hour (REF-465; transfer to a trained athlete's performance is inference); the easy volume is what you have, not what a model prescribes.

Minimum-dose work exists and is real — three 10-minute sessions a week, each holding only one or two 10–20 s all-out sprints, raised VO2max by 12–15% over 6 weeks (REF-051, B, n=29) — but that was measured in sedentary adults; treat it as a floor for a bad month, not a program, and expect far less in a trained rider.

Inference: expect a ceiling — this maintains and improves modestly, it does not build a big engine.

~5–8 h/week — the band most Blocks athletes live in. 2 quality sessions, one long-ish easy ride, remainder easy. Not from a study; the base's own practice: in practice this lands near 70/30 by TIME, not 80/20, and that is correct rather than sloppy: the hard fraction is high because the total is small, and the alternative is not more easy volume, it is fewer stimuli.

Do NOT add a third hard session to "catch up" — the recovery cost is paid from the same small budget (inference).

~8–12 h/week. The classical pattern starts to make sense. 2–3 quality sessions (base practice, not a measured optimum); an 80/20-type split by time becomes achievable (the 80/20 label is a convention, not a tested ratio — the polarized arm that won in REF-303 was ~68/6/26 by zone, full text), and the choice between pyramidal and polarized becomes real rather than theoretical.

12 h+/week. Not from a study; the base's own practice: elite templates apply as written, and the seasonal SEQUENCING evidence becomes the more interesting question than the ratio itself.

Sequencing beats the label

REF-509 (A-, RCT, n=60, 16 weeks, well-trained male runners, total load held constant) is the entry that matters more than the polarized-vs-pyramidal argument: all four arms (pyramidal, polarized, and both 8+8-week orders) improved thresholds and 5-km time, but starting pyramidal (more Z2 plus a moderate share of threshold/tempo) and shifting toward polarized (Z1 + Z3, minimum of the middle) as the event approaches improved most — ~3.0% relative VO2peak, ~1.5% 5-km TT; the reverse order did not confer the same advantage on the threshold and 5-km variables.

So the honest answer to "am I polarized or pyramidal" is "which month is it".

The effect sizes are small and it was runners; for cyclists the sequencing principle is inference. REF-303 (A-, n=48 well-trained) shows polarized (VO2peak +11.7%, ramp TTE +17.4%) beating threshold-dominant and high-volume training, neither of which produced further performance gains over 9 weeks — but that is a comparison in well-trained athletes (the entry gives no weekly hours; adequate volume is our assumption), not a licence to polarize a 5-hour week.

Two more variables — time to the goal and current fitness

Time course ([REF-465]): SIT was particularly effective for mitochondrial content and VO2max in the early stages of training; endurance training and HIT gave slower but steady improvements over more training weeks, and the paper reports no HIT plateau. Lower initial fitness meant larger percentage gains. Inference: a short runway favours intensity; a multi-year horizon favours accumulating easy hours.
A block redistributes hard sessions; it does not add them. With similar total HIT over 12 weeks, 5 HIT sessions in one week then one a week for 3 weeks raised VO2max +8.8% vs +3.7% for 2 a week (REF-136, B, n=15 trained cyclists).One-week blocks of moderate-intensity intervals or HIT, each followed by 6 days of active recovery, both raised 15-min power (+4.9% vs +2.8%, no difference), the moderate block gaining more at 4 mmol/L (REF-421, B+, crossover, n=22); against time-matched regular training the moderate block raised 4-mmol power +4.0% vs −1.3% (REF-137, B+, crossover, n=30 men).All are well-trained cyclists; a block at 5-8 h/week is untested — inference there.
More HIT is not the long-term lever in the already well-trained (REF-087, B+, narrative review): athletes training 10-13 sessions a week do about 80% of SESSIONS easy, and intensification studies show no convincing long-term gain from a greater HIT emphasis. It says nothing about beginners.
What the fast and the successful did (observational, not causal). U23 cyclists who reached elite level had logged more hours, distance and sessions and also had higher 2-12-min and critical power; the intensity split is not reported (REF-501, B, n=20).In 14 World Tour women over 43 seasons, the successful seasons had more distance, duration and kilojoules and more time in power zones 1-2 at a similar mean power (REF-502, B).In runners, pyramidal and polarized distributions were more effective than threshold-heavy training across 16 studies, only 6 of them interventions (REF-773, B+, systematic review); highly trained and elite distance runners are typically pyramidal, with a threshold and a faster session at least weekly, shifting toward polarized in the competitive period (REF-811, B+, systematic review of 10 descriptive articles); in 8 sub-elite runners over 6 months, time below VT1 dominated (4581 vs 1354 vs 487 min) and tracked the longer cross-country race time (r = −0.97) (REF-817, B).
Keep one hard session a week through a base phase. Not from a study; the base's own practice.

Practical Application

Under ~5 h/week count sessions, not percentages (2 quality + whatever easy fits); at 5-8 h, 2 quality sessions, one long-ish easy ride, the rest easy — ~70/30 by time, not 80/20; at 8-12 h, 2-3 quality sessions and 80/20 becomes real. Not from a study; the base's own practice — the hour bands and session counts are house conventions built on the per-hour and frequency findings of REF-465.
Ask for real hours before proposing any distribution, and take the number the athlete actually rides, not the one they intend to.
Prescribe in sessions below 8 h, in percentages above it. "Two quality sessions a week" is executable; "80/20" at five hours is a rounding error the athlete cannot act on.
Never add hard sessions to compensate for missing hours. The limit on quality is recovery, and recovery does not scale with ambition (inference).
Match the phase, not the fashion: pyramidal early, polarized late (REF-509).
Keep the easy easy. The most common failure at low volume is that the "easy" hours drift into the middle, which (inference) costs the hard days their quality and buys little aerobically — the nearest evidence is REF-303, where a threshold-dominant distribution (~46/54/0 by zone) made no further gains in any endurance-performance variable apart from a slight work-economy improvement; drift at low volume was not itself studied.
Sprints inside easy rides are cheap (inference: fatigue was not measured) and, by inference from one transition-period study in elite men, worth keeping through low-volume and transition phases — one weekly set of 3 × 3 × 30 s sprints inside an easy ride during a 3-week transition left elite cyclists +7.3% on 20-min power six weeks later vs −1.3% for easy-only (REF-165, n=11); in trained athletes, near-maximal speed-endurance training maintained oxidative capacity and improved 30 s–4 min performance despite reduced total volume (REF-695, B, narrative review). *Inference:* such work holds high-intensity tolerance when total volume is cut.

What NOT to do

Do not quote 80/20 to an athlete whose weekly total makes it arithmetic rather than training.
Do not scale an elite template by dividing. Inference: the elite distribution is a consequence of volume, not a cause of performance.
Do not treat the minimum-dose sprint studies (REF-051, B, sedentary sample) as a program. They answer "what if there is almost no time", not "what is best".

Limitations

No RCT compares distributions at matched LOW volume. The 5–8 h guidance here is an inference from the dose-response and session-count evidence, not a direct finding — it is the framework's weakest joint and should be presented to the athlete as reasoning, not as proof.
Hour bands are conventions, not thresholds; a 7-hour week of an experienced Tier 3 rider is not the 7-hour week of a beginner.
The elite distribution studies are observational training-log analyses — they describe what winners did, which is not the same as what caused winning. REF-197 in particular is a Strava dataset: its zones are assumed from critical speed, not measured.
Nothing here accounts for life stress, which at 5–8 h/week is often the real constraint on how much quality the athlete can absorb.
How this entry was checked and changed
Why this level

(the hour bands and session counts — the decision itself — are base practice and inference; the surrounding evidence is REF-465 A, REF-509 and REF-303 A-, block and elite-distribution entries B/B+)

Content correction 2026-09-27
"1–3 HIT sessions per week is the best return per hour (REF-465)" -> REF-465 gives per-hour ratios (HIT ~1.7×, SIT ~3.9× ET for mitochondria) and frequency 6 > 4 > 2; it does not state a sessions-per-week optimum. Session counts relabelled as house practice.
"sub-2:15 marathoners at 78/8/14 (REF-197)" -> abstract: 119,452 runners, fastest group >3× volume, more Z1, >80% pyramidal (not polarized, no 78/8/14).
"3×20 s all-out per week (REF-051)" -> abstract: 3 × 10-min sessions/week with 1–2 × 10–20 s sprints, 29 sedentary adults, VO2max +12–15%.
REF-303 "polarized beating chronic threshold" kept, numbers added (VO2peak +11.7%, TTE +17.4%); its POL split was ~68/6/26, so "80/20 becomes achievable" labelled a convention.
REF-509: added that all four arms improved and PYR→POL was largest (~3.0% VO2peak, ~1.5% 5-km), runners only.
REF-165 / REF-695 claims given their actual numbers (+7.3% vs −1.3% 20-min power; maintenance under reduced volume); 70/30, hour bands and "2 quality sessions" labelled practice/inference.
Sync after the full-base pass: REF-650 is Tier 3-4 "highly trained" swimmers (not elite Tier 4-5) and its event split has three bands (sprint polarized/threshold, middle-distance threshold/pyramidal, distance pyramidal); wording and tier ranges updated. REF-638 (now B, n=36 juniors) — the 95% below 2 mmol/L claim still matches.
REF-303 level A -> A-: that entry was regraded (parallel RCT, n=48 in four arms).
Residual sync (REF-670/695 re-checked): REF-670 claim matches (shift toward zone 1 over 38 years). "REF-695 speed-endurance maintains short-duration performance when volume is cut" -> in trained athletes it maintained oxidative capacity and IMPROVED 30 s–4 min performance despite reduced volume (B, narrative review); "holds high-intensity tolerance" labelled inference.
Level parentheticals synced to the cited entries' current Level lines: REF-509 A -> A-.
Resync 2026-09-28: Lead "the 80/20 template ... was measured on people training 15–25 h/week" -> none of the cited cohorts reports a weekly-hours figure; restated as elite, high-volume cohorts with the 15–25 h labelled the base's reading (as the body already said).
Resync 2026-09-28: Level parenthetical "synthesis over A-level primaries" -> REF-465 is A; REF-509 and REF-303 are A-; the elite-distribution entries (REF-638, REF-670 B; REF-650, REF-197 B+) are lower.
Resync 2026-09-28: "REF-303: the threshold-dominant group made no further gains in any endurance-performance variable" as proof that easy-hour drift 'buys nothing aerobically' -> REF-303: THR (~46/54/0) made no further gains apart from a slight work-economy improvement; the drift claim labelled inference.
Resync 2026-09-28: REF-509 "the reverse order did not confer the same advantage" -> scoped to the threshold and 5-km variables, as REF-509 states (POL→PYR matched PYR→POL on relative VO2peak in its full-text table).
Resync 2026-09-28: Cross-reference "REF-638 REF-650 REF-670 REF-197 ... Tier 3-5 evidence" -> REF-197 is mixed Tier 1-5 (119,452 runners across performance levels); noted.
Revision 2026-10-01: merged SYNTH-19 in as "Two more variables — time to the goal and current fitness".Added the REF-465 time course (SIT early, ET and HIT steady, no HIT plateau; lower fitness gains more), the block evidence (REF-136, REF-421, REF-137), REF-087 (80% of sessions; more HIT gives no convincing long-term gain in the well-trained), the observational pair (REF-501, REF-502), three runner entries this card did not cite (REF-773, REF-811, REF-817) and the weekly hard touchpoint in base as labelled practice.
Revision 2026-10-01: contradiction with SYNTH-19 on hard sessions per week resolved in favour of this card. SYNTH-19 gave 2-3 HIT at ≤ 8 h and 2 at 10-15 h (more hard sessions at fewer hours); this card gives 2 at 5-8 h and 2-3 at 8-12 h.Both are house practice — REF-465 gives per-hour efficiency and a frequency effect, not a tolerable session count — so the old line has no stronger evidence and its matrix (hour bands ≤8 / 8-15 / ≥15, TID percentages, block rhythm, VO2max bands) is not moved. Also not moved: the gross-efficiency lines (REF-488, REF-485) — they belong to SYNTH-17 / SYNTH-55.
Verification 2026-10-01: the section merged on 2026-10-01 ('Two more variables — time to the goal and current fitness') was checked line by line against REF-465, 136, 421, 137, 087, 501, 502, 773, 811, 817 — every number, n, design and level matches the entry; populations are stated; the one prescriptive line (one hard session through base) is labelled practice; the block bullet carries its own 'untested at 5-8 h — inference' clause. No finding in the merged section.
Verification 2026-10-01: all other cited numbers match (REF-051, 165, 197, 303, 509, 638, 650, 670, 695). REF-509 is cited as 'RCT' as in its head block; the entry's Protocol notes the allocation method is not stated in the abstract.
Verification 2026-10-01: if the Level is lowered to B, the 'SYNTH-26, B+' parentheticals in SYNTH-27 and SYNTH-03 need the same change.
Verification 2026-10-01: internal tension noted, not separately fixed — the lead dates the elite templates to ~15–25 h/week while the 12 h+ row says they 'apply as written'; and the newly cited REF-087 gives 80/20 by SESSIONS (10-13 a week), whereas the 8-12 h row and 'prescribe in percentages above 8 h' use 80/20 by time.