SYNTH-76 · Guide
Planning the season
No periodisation model wins, so count back from one to three A-races, keep a short off-season that never drops intensity to zero, move from general to specific as the race nears, and between two close races recover, maintain and sharpen instead of building; season length, the number of peaks and the close-race window are practice, not evidence
The question it answers
"How do I lay out my year?" — the off-season's length and content, the order of phases, how many races can be peaked for, the gap between two races 3-4 weeks apart, a mid-season break, and planning with no race or few hours.
Not repeated here: what a block is for and the horizon bands SYNTH-42; taper SYNTH-34; ramp and deload SYNTH-29; hours and intensity distribution SYNTH-26; return after a break SYNTH-36; recovery after a race SYNTH-70; the gym SYNTH-43 / SYNTH-44; masters (same plan, wider gaps, intensity kept) SYNTH-35.
Default (what to prescribe first)
For a healthy Tier 1-3 athlete with a dated goal:
1
Rank the races, then count back from the A-race. Not from a study; the base's own practice: 1-3 A-races a season, B-races get a few easy days, C-races are trained through. The horizon bands that say what each stretch is for are SYNTH-42 (house defaults).
2
Order (practice): off-season → general work → specific work → taper → race → recovery. In line with it, twelve Norwegian world-class coaches all describe traditional periodisation, with volume falling and competition-specific training rising as the competitive period approaches (REF-138, B, coach descriptions, Tier 5 — transfer, outcomes not tested).
3
Do not sell a model. In 30 trained cyclists, 12 load-matched weeks of block and traditional periodisation gave the same time-trial gains (REF-098, B+); a critical review argues that the named models share untested planning assumptions (REF-945, C+, no data). Pick the structure the athlete can execute.
4
Off-season: short, and never intensity-free. Not from a study; the base's own practice: up to ~7-10 days off or unstructured, then 2-4 weeks of reduced volume with one hard session every 7-10 days and the gym block started.
5
A full taper only for A-races (practice); its shape and the three-week limit are SYNTH-34.
Decision
| Situation | Do this | Source and how sure |
|---|---|---|
| Off-season: how long | Practice: 2-4 weeks reduced, at most ~7-10 days fully off | With under 4 weeks of insufficient stimulus highly trained athletes rapidly lose VO2max and blood volume, and capillary density and oxidative enzymes are already reduced, while strength declines little (REF-374, B+, narrative review, 2000). Complete cessation: VO2max −7% in 21 days, −16% by day 56 (REF-377, B, n=7). Stopping vs reducing, randomised: in 14 top-level male kayakers 5 weeks of complete cessation cost 10.1% of VO2max, reduced training 4.8% (REF-950, B, 7 per group, kayak — transfer; content of the reduced training not given). In 21 recreational runners held to <2 h/wk for 8 wk after a marathon, plasma volume fell 6.0% and left ventricular mass 10.3% within 4 wk (REF-942, B, uncontrolled). The ~10 days is SYNTH-36's convention; no entry compares off-season lengths |
| Off-season: what stays | One hard session every 7-10 days, or one sprint set inside an easy ride each week | One HIT session every 7-10 days in an 8-week transition vs low intensity only: power at 4 mmol/L 10.6% (±8.2%) and 40-min power 12.4% (±7.6%) better between groups afterwards (REF-166, B, 13 well-trained cyclists, not randomised). One weekly session of 3 × 3 × 30-s sprints in a 3-week transition with a 64% load cut: 20-min power +7.3% vs −1.3% six weeks into the next preparation (REF-165, B, 11 elite men, not randomised; both trials are in cyclists — transfer to runners and triathletes). Losses are limited if intensity is maintained, frequency reduced only moderately and volume markedly (REF-374). The reverse fails: with frequency and duration unchanged, cutting work rates by one-third or two-thirds for 15 wk lost VO2max and 21% and 30% of long-term endurance (REF-951, B, n = 12, trained for 10 wk by the study). In general populations endurance held for up to 15 weeks on 2 sessions a week or 33-66% less volume when intensity was kept (REF-587, B+, narrative review; athlete data insufficient) |
| Off-season: cross-training | Welcome, but keep some of the athlete's own sport | REF-374: athletes should use similar-mode exercise; moderately trained people can also benefit from dissimilar modes. In 30 well-trained runners, 8 weeks of extra training done as swimming improved a 3.2 km time trial by 13.2 s, done as running by 26.4 s (REF-952, B+, randomised; whether the two differed is not stated). Alternating run and cycle sessions for 5 wk improved 5000-m time as much as running only (N = 6 and N = 5; REF-953, B). In 14 trained cyclists gross efficiency fell from 20.3% to 19.4% after the season (REF-487, B, observational); keeping sport-specific volume to defend it is the base's own practice |
| Off-season: strength | Start the heavy block here; never zero later | Dose and block length: SYNTH-43 (cyclists), SYNTH-44 (runners). Elite cyclists who stopped lifting for the first 8 weeks of racing lost squat jump, Wmax and 30-s Wingate power vs controls (REF-744, B, 7 per group); one session a week kept the strength gains through 13 weeks of racing (REF-800, B, 6 per group, well-trained cyclists) |
| Block or traditional | Either; a block is a redistribution of the same hard work | No outcome basis for choosing over 8-12 weeks in trained road cyclists (REF-103, B+, systematic review, seven small studies). Pooled over six studies in trained to well-trained athletes, block had small advantages in VO2max and Wmax (SMD 0.40 and 0.28), both intervals starting just above zero, mean PEDro 3.7/10 (REF-946, A-) — an option, not a proven winner. One small trial favoured the block: VO2max +8.8% vs +3.7% with similar total HIT (REF-136, B, n=15). Order of interval formats made no difference in 63 trained cyclists (REF-096, A-) |
| Intensity mix far → near | More threshold share early, more polarised as the race nears; all orders worked | In 60 well-trained male runners over 16 weeks pyramidal-then-polarised improved most (~1.5% 5-km time), every arm improved, and the practical gap is modest (REF-509, A-; men only; transfer to cyclists and triathletes). The same shift is the authors' reading of elite runners' practice, not a tested effect (REF-811, B+, Tier 3-5) |
| Reverse periodisation (intensity first, volume later) | Acceptable when the calendar or the weather forces it; no performance reason to prefer it | Across 11 studies and 200 athletes it was no more effective than traditional or block periodisation for swimming, running, muscular endurance, maximum strength or maximal oxygen uptake (REF-948, B+, systematic review, no pooling, two randomised trials, no cycling result named): no demonstrated advantage, not proven equal |
| How many peaks | Practice: 1-3 A-races; distinct peaks ≥ ~8 weeks apart, or two races within ~4 weeks treated as one peak | Multiple peaking was essentially unstudied in 2009 (REF-109, C+, narrative review). The 8-week figure is the base's own derivation from SYNTH-42's horizon bands (house defaults; they concern the block's lever, not peak spacing) |
| B- and C-races | B: a few easy days before, no real taper. C: trained through, counted as that week's intensity | Practice. In two top-5 Tour de France riders high-intensity volume rose in race weeks — races carried the hard work (REF-163, B, n=2, Tier 5 — transfer). Tapers of ≤7 days also improved time trials (REF-3492, A, 14 studies) |
| Two A-races 3-4 weeks apart | Recover → maintain → sharpen; no new build (below). At 1-2 weeks apart: recover, one short race-intensity session, race | Practice assembled from SYNTH-70, REF-374, REF-3492; no entry tests it |
| Mid-season break | Practice: 5-7 easy days after the first peak when the season runs long or signals say so | No entry tests it. Elite male road cyclists finished a season with haemoglobin down (~15.4 → ~14.7 g/dL) and high CK and urea (REF-095, B, Tier 4 — transfer). SYNTH-29 triggers a deload on signals, not on the calendar; the 'season runs long' trigger here is practice |
| No race (fitness goal) | Replace the race with dated tests and 3-6-week blocks | Practice on SYNTH-42's block thesis. Goal-setting improved performance by 0.34 SD, moderate and combined short- plus long-term goals best (REF-631, A, 36 older studies, often lab or short motor tasks — transfer) |
| Limited hours (under ~8 h) | Phases change the CONTENT of the two quality sessions, not the hours | Inference from SYNTH-26 (sessions, not percentages, at low volume; house bands) |
The 3-4-week window between two A-races. Not from a study; the base's own practice:
Week 1 — recover. Easy only, length set by the first race (SYNTH-70: ~48-72 h after a cycling race, ~2 weeks after a marathon — both practice). Easy running from 48 h after a marathon did not slow muscle-damage recovery (REF-761, B+, 64 recreational runners). After a marathon or ultra the window is mostly recovery: the second race is run on the first build.
Middle 1-2 weeks — maintain. Return to normal volume with one or two key sessions at loads the athlete has already done. Why not stay tapered: tapers of 22 days or more were the smallest and not significant (REF-3492), and fitness is lost within four weeks of insufficient stimulus (REF-374) — inference from both, not a tested rule.
Last 5-10 days — sharpen. Volume down, intensity and frequency kept (SYNTH-34).
No new build. Inference: there is no time to absorb a new stimulus and taper it; a new lever is not started this close to a race (SYNTH-42, house default).
How to progress / periodise
General phase (far out): the missing stimulus and volume (SYNTH-42, SYNTH-26), heavy strength twice a week (SYNTH-43, SYNTH-44). In elite male road cyclists a winter mesocycle with 77% of training below VT1 raised MLSS power from ~225 to ~250 W and the later intensity mesocycle added little (~255 W) (REF-095, observational, no control — transfer).
Develop or maintain, per session (Olympiatoppen guideline, level C — expert practice, not a study): for a high-level cyclist the guideline suggests per-session doses to develop versus to maintain — I-1 ~180-420 vs ~120-240 min, I-3 ~60-90 vs ~45 min, I-4 ~40-60 vs ~30 min, I-5 ~30 vs ~15 min, I-6 ~15 vs ~6-8 min (interval zones as total interval time; REF-1077, C). Inference: a quality that is not this phase's target can be held on the maintenance dose.
Plan the supporting pieces by phase too: a review argues that recovery interventions should be periodised — withheld or emphasised by phase (REF-949, B+, narrative, Tier 4-5; no doses or timings).
Weight loss belongs in the off-season, not in a build (SYNTH-54; the timing is practice).
When NOT to use it / red flags
A year with no off-season. Rest is how a year stays under the energy ceiling (SYNTH-24, B, Tier 4-5 data).
A new block between two close A-races, or a new lever started under ~8 weeks out (SYNTH-42, house default).
Dropping the gym when racing starts (REF-744, elite men).
Tier 1 without a consistent base: the season plan is regularity (SYNTH-42 gate).
What the evidence does and does not show
It shows that over 12-16 weeks, at equal load, block versus traditional structure and the order of interval formats change little in trained cyclists (B+/A-), that pyramidal-then-polarised gave a small edge in well-trained male runners (A-, one trial), that block has at most a small pooled edge and reverse periodisation none (A-, B+; small studies), that reduced training holds fitness only if intensity is kept (B+ narrative review; small B trials) and loses less than a full stop (B, kayakers), and that strength gains decay within two months of stopping (B, elite men).
It does not show how long an off-season should be, how many peaks a season holds, what to do between two close races, whether a mid-season break helps, or that any season layout beats another over a full year in amateurs, women or triathletes — those rows are coaching practice.
Evidence gaps
Reverse periodisation in cyclists, and over a full season: REF-948 names no cycling result and holds only two randomised trials.
Off-season length: no entry compares lengths; full rest vs reduced training is tested once, in kayakers, with no follow-up into the next season (REF-950).
Mid-season break; women; triathlon seasons; athletes under 6 h a week: nothing.
How this entry was checked and changed
Content correction 2026-09-27
Revision 2026-10-01: folded REF-946 (Mølmen 2019 meta-analysis: small pooled edge for block in VO2max and Wmax, six low-quality studies) and REF-945 (Kiely 2012, critique, no data). The default 'do not sell a model' stands; the block row now says block is an option with a small uncertain edge rather than 'no outcome basis' alone.
Revision 2026-10-01: folded REF-948 (González-Ravé 2022). The reverse-periodisation row changed from 'no basis to recommend or to forbid, no trial in the base' to 'acceptable when the calendar forces it, no performance reason to prefer it' — 11 studies, 200 athletes, no advantage over traditional or block.
Revision 2026-10-01: off-season rows now carry REF-950 (kayakers, randomised: complete cessation −10.1% VO2max vs −4.8% with reduced training over 5 weeks), REF-942 (recreational runners after a marathon, uncontrolled), REF-951 (Hickson 1985: frequency and duration held, intensity cut — fitness lost) and REF-952/953 (cross-training: added swimming helped about half as much as added running; alternating run/cycle matched run-only over 5 wk). 'Short, and never intensity-free' keeps its wording; its length is still practice.REF-949 (Mujika 2018) added as one line on planning recovery by phase; it does not fill the multiple-peaking gap.
Revision 2026-10-01: six Evidence-gap lines narrowed. Level unchanged at B-: off-season length, number of peaks, the close-race window and the mid-season break are still house practice.
Verification 2026-10-01: all 33 cited entries opened (REF-095, 096, 098, 103, 109, 136, 138, 163, 164, 165, 166, 206, 3492, 374, 377, 468, 487, 509, 587, 631, 744, 761, 800, 811, 942, 945, 946, 948, 949, 950, 951, 952, 953) plus the lines of SYNTH-24, 26, 29, 34, 36, 42, 43, 54, 70 it points to.Every number matches its entry (-7% / -16%; 10.1% vs 4.8%; 6.0% and 10.3%; 10.6% ±8.2 and 12.4% ±7.6; +7.3% vs -1.3%, 64%; 21% and 30%; 13.2 s vs 26.4 s; 20.3% -> 19.4%; SMD 0.40 / 0.28, PEDro 3.7; +8.8% vs +3.7%; ~1.5%; 11 studies / 200 athletes; 0.34 SD; 15.4 -> 14.7 g/dL; 225 -> 250 -> 255 W; 9%). Levels, designs and n next to each refId match the head blocks.No NUMBER or WRONG finding.
Verification 2026-10-01: Level B- is honest. Decision rows: 5 rest on cited trials or reviews (what stays, strength, block vs traditional, intensity mix, reverse), 2 mix a practice number with evidence on direction (off-season length, cross-training), 6 are labelled practice (peaks, B/C races, close A-races, mid-season break, no race, limited hours). The title and the closing paragraph say the same.
Verification 2026-10-01: source-status check — REF-095 and REF-950 have no DOI but carry a verified PMID; REF-744 has a non-standard head block but a DOI; REF-374's entry says 'do not quote numbers from this entry' and the SYNTH quotes none from it. No SOURCE finding.
Verification 2026-10-01: not flagged in the SYNTH but worth knowing — REF-374 (cited) says capillary density is 'reduced already within < 4 weeks', while REF-377 (also cited, for VO2max only) found capillarisation 'did not decline' over 84 days. The SYNTH quotes the REF-374 line only; the two entries disagree on capillaries.
2026-10-01: level tags next to cited frameworks updated to their current levels after the independent audit.
2026-10-02: aligned with the rewritten SYNTH-42 (missing stimulus, one lever per block, level C+): four "limiter" wordings -> "the block's lever" / "a new lever is not started" / "the missing stimulus"; no rule changed.
Revision 2026-10-04: added Olympiatoppen's Intensity Scale v2 (2024) REF-1077 as practice lines labelled "(Olympiatoppen guideline, level C — expert practice, not a study)"; no default, decision row or level changed.