SYNTH-17 · Guide
Efficiency and economy
They rise as a by-product of hard training, heavy lifting and, probably and slowly, years in the sport, not from cadence coaching, pedalling drills or a low-carbohydrate diet; read them within one lab and one protocol, and expect them to fall late in long efforts
The question it answers
"Can I become more efficient, and how?" What raises cycling GE or running economy, what does not, how to read a lab value, and what to tell an athlete who wants efficiency drills. Doses are owned elsewhere: gym for cyclists SYNTH-43, for runners SYNTH-44, jumps SYNTH-45, cadence SYNTH-55, durability SYNTH-32. Speed per watt (position, equipment) is a different question: SYNTH-48, SYNTH-62.
Default (what to prescribe first)
1
**Do not programme "efficiency sessions".
Keep the hard sessions and the gym, and efficiency follows.** In 29 competitive male cyclists, GE rose only in the 6 weeks that included prescribed high-intensity sessions (+1.6% and +1.4%) and did not change in the intensity-restricted group (+0.1%), nor further in the first group over the following 6 weeks (+0.4%) (REF-488, B+, randomised).
A matched high-volume arm was not tested, so volume is not ruled out. Across one season in 14 trained cyclists, GE went 19.6% → 20.6% in the build-up and back to 19.4% afterwards, and the gain tracked total training time (r = 0.84) and time above OBLA (r = 0.80) (REF-487, B, observational).
2
Lift heavy. In 17 controlled trials of 262 cyclists, heavy strength training (≥80% 1RM) improved cycling efficiency (ES 0.35) without changing VO2max (REF-743, A, GRADE low).
In runners, heavy loads improved economy by a small amount (ES −0.266) and combined methods by a moderate amount (ES −0.426), while submaximal loads (40-79% 1RM) and isometric work did not significantly (REF-801, A-; isometric: three small calf-only studies, CI crossing zero). Programmes: SYNTH-43, SYNTH-44.
3
Promise a small change over a block and say so. GE has a test-retest CV of 4.2% (REF-489, B, n=14).
Decision (what moves it, what does not)
| Lever | Verdict | Source and how sure |
|---|---|---|
| High-intensity sessions within normal training | Moves cycling GE | REF-488 B+ (competitive men) and REF-487 B (trained cyclists, sex not stated in the entry). How much intensity a rider should carry is not answered by these studies |
| Heavy strength, cyclists | Moves it | REF-743 A. Single trials: economy +4.8% after 8 weeks of 4×4RM half-squats in 13 competitive cyclists (REF-078, B); +8% with bilateral lifting but not with alternating unilateral in a pilot of 24 trained triathletes/cyclists (REF-148, B+); improved in 19 well-trained women, no size in the abstract (REF-073, B). In 14 young elite cyclists economy improved in the endurance-only group instead (REF-075, B), so the effect is not found every time |
| Heavy strength, runners | Moves it | Energy cost of locomotion improved across running, cycling, skiing and swimming (SMD 0.65; REF-074, A-). Economy generally improved 2-8% in a review of 24 studies of distance runners (economy measured in 20; heavy, explosive and plyometric programmes not separated), a spread and not in every study (REF-072, A-). In 16 master marathoners only maximal-strength training improved economy at marathon pace (+6.17%); general resistance training did not (REF-864, B, preliminary) |
| Masters | Same levers | Three weeks of knee-extension training raised cycling delta efficiency in 9 masters and closed the gap to young athletes (REF-860, B, no non-lifting control). Training status and age both affected GE in 40 men (REF-485, B, cross-sectional) |
| Plyometrics | Runners: small. Cyclists: no trial | Oxygen cost at 14 km/h −1.96% at p = 0.05 in track-and-field athletes (REF-518, A). In runners the pooled gain appeared only at ≤12 km/h (REF-801), while one trial in 15 highly trained runners found it only at 18 km/h (REF-868, B). Do not promise a gain at a given pace. Details: SYNTH-45 |
| Years of training | Probably, slowly | A review concludes well-controlled studies show training raises GE, with mechanisms undetermined (REF-076, B+, narrative). One Tour de France champion gained 8% in muscular efficiency from age 21 to 28 (REF-077, C+; a single case, and the athlete later admitted doping over that period) |
| Endurance training itself, runners (volume, years, intervals) | On the map; size not given | A narrative review lists endurance training — with training history and training volume as suggested factors — and uphill and level-ground high-intensity interval training among the strategies studied for running economy, beside resistance and plyometric training, altitude, flexibility, nitrate and caffeine (REF-977, C+, narrative review; no effect sizes, doses or ranking in its abstract — a map of levers, not evidence of size) |
| Imposed low cadence at moderate intensity | Does not move it | 12 weeks of 40-rpm intervals at 73-82% HRmax changed neither GE nor performance in 22 veteran cyclists (REF-118, B+). No moderate-intensity grinding: SYNTH-55 |
| Riding at the "most economical" cadence | Not a target | The most economical cadence was 53.3-59.9 rpm in every group, yet cyclists and equally fit runners chose 90-100 rpm (REF-881, B, men). In another study metabolic cost was lowest at 60 rpm while riders chose 81 rpm, and the chosen cadence gave the highest vastus lateralis power capacity with no significant efficiency penalty (REF-238, B) |
| High-cadence intervals | Cadence-specific only | GE rose at 90 and 110 rpm (P = 0.05) after 6 weeks, but the low-cadence group gained more on the time trial (REF-259, B+, 16 men) |
| Pedalling-technique cues, pedal-force feedback | Do not cue | Evenness of power around the crank correlated with GE (r = 0.75) and force effectiveness did not predict it; cross-sectional, so no gain can be promised from drills (REF-739, B, n=21). A composite entry reports that consciously pulling up lowered GE (REF-499, C+, numbers unverified) |
| Indoor trainer | Neutral | Muscle recruitment changed, efficiency did not (18.5 vs 18.8%; REF-254, B+, 19 men) |
| Chronic low-carbohydrate, high-fat diet | No economy gain at race pace (the other diets gained) | In 29 world-class race walkers over 3 weeks, oxygen cost at race pace fell in the high-carbohydrate and periodised groups and not in the ketogenic group, and 10-km time improved 6.6% and 5.3% in those two groups but not in the ketogenic group (−1.6%) (REF-486, B+; race walking, transfer to cycling and running is inference) |
| Sleep-low periodisation | One positive trial | Cycling delta efficiency +11 ± 15% vs +1.4 ± 9.3% in 21 trained triathletes over 3 weeks at the same daily carbohydrate (REF-003, B+, randomised per the full text (corrected 2026-10-12)). Where train-low fits: SYNTH-69 |
| Shoes (runners) | Moves it, by equipment | Plated shoes lowered metabolic demand ~2-3% (mean −2.75%) across 14 crossover trials, with no performance outcome (REF-720, A) |
| Muscle fibre type | Not a prescription | % Type I fibres correlated with GE (r = 0.75) in 19 highly trained cyclists (REF-484, B); a later study of 40 men found no relationship (REF-485) |
| Heat | Unknown | No entry measures GE or running economy in the heat. Hot-day pacing: SYNTH-10 |
Reading a number
Compare only within one athlete, one lab, one equation. Four energy-expenditure equations gave different absolute GE (all P < .001) with the same reliability (REF-489). Inference from that entry: a change of about one GE point on a ~20% baseline sits near the noise of a single test.
A flat VO2max is not a failed block. In trained cyclists training raised GE but not VO2max, and within riders the two moved inversely (r = −0.32 over a season; r = −0.78 over six high-intensity weeks) (REF-444, B+, n=14 and n=29). The entry notes the correlation may partly be measurement artefact.
No lab: not from a study; the base's own practice: repeat one fixed submaximal test (power or pace at a fixed heart rate) in similar conditions and read the trend monthly. It does not separate efficiency from fitness.
How to progress / periodise
It falls during long efforts. GE fell after ~60 min at ~60% VO2peak, and the fall tracked the loss of 5-min power (r = 0.91) (REF-423, B). After 2 h, part of the ~10% fall in VT1 power was lost efficiency (−8 ± 10 W) (REF-179, B, n=14).Over a 90-min run, economy worsened 3.1% in runners doing regular long runs against 6.0% in matched runners without them (REF-160, B, 26 well-trained men, cross-sectional). In 10 young professionals, fresh GE at 300 W correlated with how well CP was held after hard work (r = 0.869) (REF-424, B, correlational). Training for this: SYNTH-32.
When NOT to use it / red flags
Do not quote a percentage gain to an athlete. The single-trial figures (+4.8%, +8%, +6.17%) come from small groups.
Do not sell jumps to a cyclist as an efficiency tool. The base holds no cycling plyometric trial (SYNTH-45).
Equipment ordering is not evidence. Not from a study; the base's own practice: for speed per watt, position first, then clothing and helmet, then wheels, then frame; tyres and pressure first on rough surfaces; system mass on long climbs. No entry ranks these.
What the evidence does and does not show
Shows: heavy strength improves efficiency or economy in cyclists and runners (A/A-, small effects, low to moderate certainty); high-intensity training raised GE in one randomised trial of male cyclists (B+).
Evidence gaps
Heat and efficiency or running economy. No entry.
Volume versus intensity for GE with a matched high-volume arm (REF-488 had none).
Running economy and endurance training itself (intervals, volume, years), measured. The base's runner evidence with effect sizes is almost all strength and plyometrics; REF-977 only lists these levers (C+ narrative review, no sizes).
Cadence-variation drills and GE. No trial.
How this entry was checked and changed
Why this level
(strength → economy rests on A/A- meta-analyses with low-to-moderate certainty; training → GE on one B+ randomised trial in male cyclists and B observational work in trained cyclists; diet on one B+ trial in race walkers; technique on B cross-sectional work and a C+ composite; years on a single case)
Content correction 2026-09-27
Strength row "+2-5% economy/GE, 8-12 weeks, REF-073/075/076" -> REF-073's abstract gives direction only (economy improved, P < 0.05, 11 weeks); REF-075's abstract says economy improved in the endurance-only group while strength raised 45-min capacity 8% (16 weeks); no percentage for strength → economy in any cited abstract.
"+5-10% relative GE trained vs recreational" -> not in any entry; REF-076's abstract says cross-sectional comparisons found no GE difference and that well-controlled training studies do raise GE; the REF-076 entry's 22% vs 19% / 70% vs 40% figures are flagged. REF-077's +8% kept (abstract: 8% improvement in muscular efficiency, ages 21-28).
"LCHF impairs economy by ~5-8%" -> REF-486's abstract: O₂ cost at race pace fell in HCHO/PCHO (90% CI −7.0 to −2.6% and −5.2 to −0.9%) and did not in LCHF; 10-km +6.6% / +5.3% vs −1.6%.
"Hopker 2018" -> Hopker 2012 (REF-444) with the inverse-correlation values; REF-424 GE-durability link stated per its abstract (strong positive correlation, n=10).
"Running economy varies 30-40%, cycling GE 20-30%" -> untraced, replaced with REF-484's measured range (DE 18.3-25.6%, r = 0.75 with GE). "In-race CHO 100-120 g/h" and "60-120 g/h" -> REF-066's 30-90 g/h by duration; higher intakes labelled house practice. Cadence protocol, lab-load convention and fixed-HR surrogate labelled house practice.
Full-base sync (later 2026-09-27): REF-488 "both low + high intensity improve GE, equivalent, combination wins" -> reversed: GE rose only after the high-intensity block (+1.6% / +1.4%), not with intensity restricted (+0.1% / +0.4%).
REF-485 "training status, not age, separates GE; GE preserved into masters age" -> both age and training status had significant effects on GE (P < 0.01).
REF-490 "aero dominates on the flat", "gravity dominates on steeper gradients", "rolling resistance elevated on gravel" -> not in that review's abstract; event-type ordering relabelled as the base's own practice.
REF-073 "leg press 4-6RM, 2×/wk" -> not in the abstract; removed. REF-066 ">2.5 h: 75-90 g/h" -> "up to 90 g/h". REF-486 population (elite race walkers) and REF-489 CV 4.2% added; REF-499 flagged as C+ composite.
Level B+ -> B: REF-459 is now B+ (was A), and the efficiency primaries are mostly B or C+.
Stale "pending correction" notes on REF-424, REF-075 and REF-076 replaced: those entries are now corrected (REF-424 r = 0.869 / 0.792 with GE; REF-075 economy result added; REF-076 22%/19% removed).
Resync 2026-09-28: REF-075 "16 weeks of heavy squats" -> the entry says the strength program is not described in the abstract (the squat detail was deleted from it); now "strength training added to endurance training".
Resync 2026-09-28: REF-486 magnitude cell "protects baseline economy; LCHF erodes it" -> the entry: O₂ cost at race pace fell in HCHO/PCHO and stayed at pre-intervention level in LCHF (no gain, not a loss).
Resync 2026-09-28: REF-073 "Vikmoen 2015" -> 2016 (the entry's year); REF-487 "in-season GE gains" -> pre-competition-phase GE gains (where the entry places the r = 0.84 / 0.80).
Resync 2026-09-28: Level line "only REF-444, REF-486 and REF-488 are B+" -> REF-066 and REF-076 are also B+, and REF-460 is A- (energy systems, not efficiency); level unchanged.
2026-10-01: rewritten in the decision-framework format as a general framework on what moves cycling gross efficiency and running economy. All personal material removed (the named owner and his events in the gravel row); the framework now speaks of athletes in general.The default changed from "efficiency is the highest-leverage trait, train it with cadence drills" to "keep the hard sessions and the heavy lifting; do not programme efficiency sessions".Resolved in favour of the newer frameworks: the cadence-variation protocol (50-60 rpm blocks in Zone 3, "hypothesised +0.5-1.5%") deleted as contradicting SYNTH-55 and REF-118; "90-120 g/h for adapted athletes" and the REF-066 rates removed, in-race grams belong to SYNTH-02; strength doses left to SYNTH-43 / SYNTH-44, plyometrics to SYNTH-45; the event-type aero decision tree reduced to one labelled practice line, position and aero evidence lives in SYNTH-48 / SYNTH-62.The PMEP "compounding" principle (2% GE inference) and the stayer-phenotype inference deleted. Folded in: REF-743, REF-078, REF-148, REF-860, REF-864, REF-074, REF-072, REF-801, REF-518, REF-868, REF-720, REF-118, REF-881, REF-238, REF-259, REF-739, REF-254, REF-003, REF-092, REF-423, REF-179, REF-160, REF-845, REF-620.
Verification 2026-10-01: every refId in the body (39 entries) opened in the snapshot; all numbers in the Default, the Decision table, 'Reading a number' and 'How to progress' match the cited entries (REF-003, 073-078, 092, 118, 148, 160, 179, 238, 254, 259, 423, 424, 444, 484-490, 518, 620, 720, 739, 743, 801, 845, 860, 864, 868, 881). Levels and n next to each refId match the entries' head blocks.
Verification 2026-10-01: Level B is honest — 27 of 31 prescriptive rows carry evidence, the two practice lines (no-lab surrogate, equipment ordering) are labelled in the same sentence. No LEVEL-OVERALL finding.
Verification 2026-10-01: the worst item is the REF-488 '+0.4% with intensity restricted' mislabel; it is inherited from that entry's own Practical Application line, which contradicts its Key Findings — the REF-488 entry needs the same fix.
Verification 2026-10-01: 'No entry measures GE or running economy in the heat' checked by grep over the snapshot — nothing contradicts it. 'No cycling plyometric trial' agrees with SYNTH-45; the cadence red flag agrees with SYNTH-55 (which says 40-60 rpm).
Verification 2026-10-01: soft tension, not flagged as a contradiction — Default says 'Lift heavy' and cites REF-801 (submaximal 40-79% 1RM did not help runners), while the Masters row cites REF-860, a 10 × 10 at 70% knee-extension programme; the row does not call it heavy.
2026-10-01: title: 'years in the sport' rests on a narrative review and one case (REF-076, REF-077); the title now carries the table row's own 'probably, slowly'.
Revision 2026-10-01: REF-977 (Barnes & Kilding 2014, C+ narrative review) added as one decision row and cited only as a map of levers for running economy (endurance training, history and volume, uphill and level intervals), explicitly with no effect sizes; no default changed, no percentage attached. The Evidence-gap line is narrowed to 'measured' endurance-training effects on running economy, which the base still lacks. Level B unchanged.
Revision 2026-10-16: REF-801's submaximal and isometric economy results reworded to "not significantly" (isometric pool: three small calf-only studies, ES −0.269, CI −0.79 to 0.252; see SYNTH-88). Level unchanged.