SYNTH-56 · Guide
ERG mode or not
ERG is a compliance tool for submaximal reps, not a better stimulus; ride sprints, maximal tests, shaped VO2max reps and race rehearsal in resistance or slope mode, and no trial in the base shows either mode trains a cyclist better
The question it answers
On a smart trainer, which sessions should run in ERG (the trainer holds the target watts whatever the cadence) and which in resistance/level or slope/SIM mode (the rider produces the watts by gear and cadence, pacing to the number on screen)? What changes physiologically, what goes wrong in ERG, and what indoor-specific factors matter? Not from a study; the base's own description: these are the equipment's modes, not constructs any cited study defined.
Default (what to prescribe first)
ERG for submaximal structured reps where hitting the number is the point — endurance, tempo, sweet-spot and threshold reps below CP, programmed over-unders (low-cadence work: only inside hard reps, and all-out torque efforts are self-paced — SYNTH-55). Not from a study; the base's own practice.The physiology gives no reason to prefer ERG on these sessions either: in 8 trained men riding 1 h at 78% VO2max, ±5% swings every 5 min gave the same VO2 (3.26 vs 3.33 L/min), HR (159 vs 158), lactate (4.3 vs 4.2 mM) and RPE as constant power (REF-241, B). ERG is chosen for compliance, not for a better stimulus.
Resistance or slope mode for anything maximal or self-paced — sprints, all-out 30-s efforts, fast-start VO2max reps, 20-min/CP tests, race rehearsal. Not from a study; the base's own practice (by construction ERG sets the power, so it cannot find the athlete's maximum).
Keep cadence habitual and consistent across ERG sessions you want to compare. Fixed watts do not fix the task: the cadence with least muscle activation rose from 57 rpm at 100 W to 99 rpm at 400 W (REF-237, B), and in 10 collegiate athletes the respiratory compensation point fell from 222 W at 60 rpm to 197 W at 100 rpm while its VO2 did not differ (REF-202, B). The consistency rule itself is inference.
Decision (who → what)
| Session | Mode | Source and how sure |
|---|---|---|
| Endurance / Z2, tempo, sweet spot, sub-CP threshold reps | ERG or free — athlete's choice; ERG if they drift above target | REF-241 B (n=8, ±5% costs nothing); the mode choice is practice |
| Long threshold reps near FTP/CP | ERG acceptable, but target a few % below CP or FTP20; a fade is the domain, not weakness | REF-747 B: TTE at FTP20 on a smart trainer, median 35 min (recreational) to 51 min (pro), n=87; REF-750 B: at CP 20-41 min, HR/VO2/lactate still rising, n=7; using ERG for these reps is practice |
| Sprints ≤30 s, SIT, neuromuscular work | Resistance/level or slope mode, big gear, self-paced all-out | Practice (ERG cannot give "all-out"); REF-243 B+: self-paced 30-s and 1-min intervals raised VO2peak ~3 mL/kg/min in 82 INACTIVE adults — Tier 0, transfer is inference |
| VO2max reps with a shaped interior | Over-under: ERG can program it. Fast-start: self-paced | REF-700 B: fast-start and over-under both beat constant-speed on time ≥90% VO2peak, equal to each other (10 XC skiers — transfer to cycling is inference); mode split is practice |
| Recovery between hard reps (~2 min) | Self-selected is fine | REF-246 B+: 50 W, 50% FTP or self-selected recovery gave the same 5 × 3-min power (300.1 / 296.9 / 298.8 W), n=18 trained, Zwift (ERG use not stated) |
| Race simulation, surges, group-ride demands, pacing practice | Slope mode or outdoors; ride the variability | REF-252 B+: 15% surges every 2 min raised VO2 and lactate vs constant at the same mean (HR: no main effect in the Results (corrected 2026-10-12)), elite n=15; REF-253 B+: outdoor power SD ~2× indoor; rehearsal value is practice (REF-128, B+, theory only) |
| Maximal tests (20-min, 3 + 12 min CP) | Self-paced; repeat the same mode and setting on retest | REF-034 B: self-paced CP 265 vs constant-load CP 250 W, n=10; REF-182 B+: indoor ≠ outdoor in pros; REF-877 B: 3-min all-out end power matched CP at constant cadence but not at fixed resistance (n=12, lab) — a mode gives its own number; choosing self-paced is practice (ERG cannot find a maximum) |
| Low-cadence torque reps | ERG (holds watts while cadence drops) or a fixed hard gear for low-cadence VO2max reps; all-out torque efforts are self-paced, not ERG. Whether to do them at all: SYNTH-55 | Practice; REF-117 A- finds no strong evidence for low-cadence training, tentatively at moderate-to-maximal intensity; REF-259 B+ (16 men): low-cadence intervals improved a 15-min TT more |
Physiology the base supports
Variable power at the same mean costs more, and the rider may not feel it. In 15 elite cyclists, four 28-min lab bouts at 70% and 95% of LT with a 15% fluctuation every second minute gave significantly higher blood lactate than constant power at both intensities and higher VO2 (main effect; pairwise significant only at low intensity in the Results) (HR higher in the abstract, no main effect of power condition in the Results (corrected 2026-10-12)); whole-bout RPE at 95% LT was similar and the rise in RPE across the bout was smaller with variable power; the authors call variable power a viable alternative, at least at high intensity (REF-252, B+, Tier 4-5, acute only).
Small swings are free. ±5% every 5 min changed nothing measurable (REF-241, B, n=8 trained men, one intensity).
Imposed constant pace can cost MORE than self-pacing. In 9 well-trained men on a rowing ergometer, 5000 m at matched mean power and time: enforced constant pace raised lactate (6.19 vs 5.24 mmol/L), core temperature (38.72 vs 38.46 °C) and iEMG, with HR and VO2 unchanged (REF-240, B — rowing, transfer to ERG cycling is inference).
Indoor riding is smoother than outdoor. 20 male competitive cyclists produced the same 20-min TT mean indoors and outdoors (280 vs 284 W, r = 0.94) but within-ride SD was 33 vs 69 W (REF-253, B+).
Cadence sets the muscular task at fixed watts. At a constant 2.5 W/kg the metabolic optimum was 60 rpm yet riders chose 81 rpm, the choice that kept vastus lateralis power capacity highest without a significant efficiency penalty (REF-238, B). The energetically optimal cadence fell after 55 min of cycling in 14 experienced cyclists (REF-261, B).High cadence at high intensity impaired neuromuscular function more than low cadence, and no cadence range can be recommended (REF-255, A-, 13 studies).
The trainer changes recruitment, not efficiency. On a fixed turbo trainer vs a treadmill-mounted bike, 19 male cyclists at 150-250 W used more quadriceps and tibialis anterior and less gluteus maximus, biceps femoris and gastrocnemius; gross efficiency 18.5 vs 18.8% (REF-254, B+; trainer mode not studied).
How to progress / periodise
Not from a study; the base's own practice: base and build blocks can run most indoor structured work in ERG; in the last 4-6 weeks before a road race, crit or TT, move key sessions to slope mode or outdoors so the race's variability (REF-252, REF-253) and pacing decisions (REF-128, theory) are rehearsed.
Not from a study; the base's own practice: an athlete who only ever rides ERG should do one self-paced maximal effort per block (test or race-like rep) so power at a self-chosen pace is known — REF-034 shows the pacing modality moves CP (~6% in 10 young men).
Indoor zones for indoor, outdoor zones for outdoor: in 14 male pros CP was 19 W higher outdoors with substantial individual variation and W′ unchanged (REF-182, B+; outdoor tests on a ~2-3% climb); in 20 competitive men the 20-min means agreed (REF-253, B+). Inference: check each athlete's own gap rather than applying a group offset.
When NOT to use it / red flags
The "ERG spiral". Not from a study; the base's own practice: late in a hard rep, cadence sags, ERG raises resistance to keep the watts, torque per stroke rises and cadence sags further until the rider stalls.Inference from REF-237 (activation rises either side of the optimal cadence, and the optimum rises with power) and REF-261 (the optimum falls with fatigue); no study measured the spiral. Practice fix: shift to an easier gear and keep cadence up before it drops; if it stalls, pause ERG or switch to resistance mode, restart, and lower the target ~5% for the remaining reps.
Heat. In 10 trained men, 32 °C vs 19 °C with facing airflow of ~29-35 km/h made a ~15-min TT 15% slower (961 vs 819 s), cut exogenous glucose oxidation 14-15% and raised HR, core temperature and RPE (REF-461, B). Inference: in a warm room ERG holds watts while strain climbs, so cap by HR/RPE or cut the target.In 9 subjects cycling 2 h at 33 °C, still air and a 9.9 km/h breeze left heat storage, body temperature and RPE higher than facing airflow of 33 km/h or more; 50 km/h added nothing, and drinking more did not help once airflow was high (REF-886, B, no performance test). Inference: a trainer rider without a fan is near the still-air condition.Not from a study; the base's own practice: a cool room and a strong fan are standard indoor equipment.
Suspect data. A SWOT commentary names questionable data accuracy as a weakness of virtual platforms (REF-244, C+). Inference: check trainer power against a known power meter before building zones on it.
Do not use ERG for sprints, all-out efforts or maximal tests (practice, see Default).
What the evidence does and does not show
Shows (acute, B+/B): at the same mean power, large surges cost more metabolically than constant power in elites (REF-252), ±5% wobble costs nothing (REF-241), imposed constant pace strained rowers more than self-pacing (REF-240), indoor power is half as variable as outdoor (REF-253), and cadence changes the muscular and metabolic task at fixed watts (REF-237, REF-238, REF-202).
Does not show: that ERG training produces more or less fitness, race performance or adherence than resistance/slope-mode or outdoor training — no entry compares them over weeks. REF-243's self-paced training worked in inactive adults only, with no ERG arm.
Level: C+ because every decision row borrows acute physiology or equipment logic; nothing tested "ERG vs not" as a training intervention.
Evidence gaps
No trial comparing training adaptations (weeks) between ERG and self-paced/resistance-mode training in trained cyclists.
No measurement of the "ERG spiral" (cadence collapse under power control at fatigue) or of cadence drift across an ERG session.
No study of fan airflow on power or HR indoors; REF-886 measured heat storage and body temperature, not performance, and gives no fan setup.
No study of smart-trainer power accuracy against a reference power meter, or of sprint mechanics (flywheel inertia) on trainers.
No study of motivation, enjoyment or compliance in ERG vs free riding; REF-245 is unverifiable.
How this entry was checked and changed
Why this level
(acute crossover physiology in trained to elite cyclists, B+/B; no training trial of ERG against self-paced or resistance-mode riding; the session-to-mode mapping is house practice)
Content correction 2026-09-27
Revision 2026-09-29: folded in REF-886 (Saunders 2005, B: still air or a light breeze left heat storage, body temperature and RPE higher than ≥33 km/h facing airflow) into the heat red flag — the fan rule now has a study behind the direction, still no performance or fan-setup data; REF-877 (Wright 2017, B: 3-min end power = CP only in constant-cadence mode) into the maximal-test row. Airflow gap narrowed. Level unchanged (B-).
Verification 2026-10-01: every refId outside Cross-References and the revision log was opened in the KB.md snapshot (27 entries). All numbers match the entries (3.26/3.33 L/min, 159/158, 4.3/4.2 mM; 222/197 W; 265/250 W; 19 W; 280/284 W, SD 33/69 W; 300.1/296.9/298.8 W; 35-51 min; 20-41 min; 961/819 s; 6.19/5.24 mmol/L; 38.72/38.46 °C; 18.5/18.8%), as do all level tags.
Verification 2026-10-01: no WRONG and no NUMBER findings. The card is unusually candid that its mode rules are practice; the findings are the level that follows from that candour (C+ rather than B-), one conflict with SYNTH-55 on submaximal low-cadence reps, and rows whose practice label is missing.
Verification 2026-10-01: REF-241 has a PMID but no DOI (entry says verified against PubMed) — not flagged. REF-245 is NOT IN SEARCH; removing its sentence from 'Does not show' leaves the gap line ('REF-245 is unverifiable') which may stay or go with it.
Verification 2026-10-01: 'rowers' also appears in the lead ('self-pacing in rowers') and in the evidence list ('strained rowers'); only the physiology bullet is given a pair — the other two should be read as 'men on a rowing ergometer'.