REF-1477 · Pelland et al., 2026 · Sports Medicine · Paper · PubMed

Do more gym sets keep paying?

Yes, with diminishing returns, sharper for strength than for size (67 studies, model output; abstract only). Frequency helped strength; for size it may be negligible. Young, mostly male lifters.

Main caveat: Abstract only. No credible intervals, slopes, heterogeneity, risk of bias, or list of studies could be read.
fairly strong evidenceMeasured on non-athletes to trained athletes

The authors classified every set in 67 studies as direct or indirect, by how specific it was to the strength or size measurement, then counted indirect sets three ways: as 1 ("total"), as 0.5 ("fractional") and as 0 ("direct"). The fractional count had the strongest relative evidence and was used for the primary models, adjusted for intervention length and training status.

The posterior probability that the marginal slope of volume exceeds zero was 100% for both hypertrophy and strength, and both best-fit models suggest diminishing returns, considerably more pronounced for strength. For frequency, that probability was below 100% for hypertrophy (compatible with negligible effects) and 100% for strength, again with diminishing returns.

Our reading: useful for the shape of the curve and for counting a squat as half a set for a muscle it trains indirectly. The abstract gives no slope sizes, no credible intervals and no point where returns stop being worthwhile. Any such number from this entry would be invented.

Two numbers in the base are traced to this paper, both second-hand and neither checkable here: ACSM's reading in REF-1476 and the conference summary's about 12 sets per muscle per week in REF-1143.

Protocol

Data: n = 2058 participants in 67 studies; 79.1% male, 20.9% female; average age 25.16 ± 5.22 years. Open data and code on the Open Science Framework per the declarations (not read for this entry).
Set counting: each contributing set classified as direct or indirect, depending on its specificity to the measurement. Indirect sets quantified as 1 for "total", 0.5 for "fractional", 0 for "direct", for both weekly volume and weekly frequency.
Analysis: multi-level meta-regressions for hypertrophy and for strength, adjusted for duration of the intervention and training status. Marginal slopes reported as posterior probabilities of exceeding zero.
Search: the declarations say JCP and JFR performed the systematic search; dates and eligibility are not in the abstract.
Funding and interests: funding not stated in the abstract record. "The authors have no conflicts of interest specifically pertaining to this article"; all five "are coaches and writers in the fitness industry".

Key Results

RelationshipPosterior probability slope > 0Shape per the abstract
Volume to hypertrophy100%Rises, diminishing returns
Volume to strength100%Rises, diminishing returns considerably more pronounced
Frequency to hypertrophyless than 100%Compatible with negligible effects
Frequency to strength100%Rises, diminishing returns
Set counting: the relative evidence for the fractional quantification was strongest.
Not in the abstract: slopes, credible intervals, the volume or frequency at which returns diminish, and any per-exercise figure.
Another paper's reading (not from this abstract, unverified): REF-1476 cites this paper for diminishing returns beyond about 2-3 sets per exercise for strength and about 18-20 weekly sets for hypertrophy; REF-1143 cites it for about 12 sets per muscle per week, with diminishing efficiency above. The two readings differ and neither could be checked here.

Mechanisms

Measured: nothing mechanistic; modelled slopes on outcome measures of strength and size.
Authors' interpretation: distinguishing direct from indirect sets "appears essential" for predicting adaptations, and volume and frequency have distinct dose-response relationships with each outcome.
Inference: a 100% posterior probability says the direction is secure in the model, not that the effect is large. Diminishing returns are a feature of the fitted function, not an observed stopping point in any trial.
Inference: the fractional rule was chosen by model comparison on these data. It describes the data well; it is not a measurement that an indirect set delivers half a stimulus.

Practical Application

Do more gym sets keep paying? Yes, with diminishing returns, sharper for strength than for size (67 studies, model output; abstract only). Frequency helped strength; for size it may be negligible. Young, mostly male lifters.
Add sets last for a strength goal (inference): sharper diminishing returns for strength argue against more sets as the first move for a cyclist or runner who wants strength, once a lift has a few hard sets. No number comes from this abstract.
Count indirect work (use with care): a squat does work the hip extensors even when the log names only the squat (our example, not in the abstract). Counting indirect sets as half fitted best at group level; this is not a rule for individuals.
Do not quote a threshold from this paper. The abstract gives none; the full text was not available for this entry.
Against the base, or at least not settled: REF-583 calls the frequency effect on strength a volume artifact, and SYNTH-43 and SYNTH-44 use "with volume equated, frequency did not change strength gains" (SYNTH-21's frequency line is about hypertrophy, where this abstract agrees). This abstract reports frequency raising strength but does not say whether the frequency models hold volume constant, so the conflict cannot be judged without the full text.

Limitations

Abstract only. No credible intervals, slopes, heterogeneity, risk of bias, or list of studies could be read. Every statement here is what the abstract states or is labelled as inference.
Model output. Dose-response here is modelled and adjusted only for duration and training status in the abstract's account; the abstract does not say how far it rests on contrasts between studies rather than within them, and diminishing returns depend on the functional form that fitted best.
Who. Average age about 25, 79.1% male, and no endurance athletes named. Nothing here accounts for the fatigue of endurance training around the lifting.
Frequency. The frequency slope may carry volume with it; the abstract does not settle this.
Interests. All authors are coaches and writers in the fitness industry, as printed. Funding not stated in the abstract record.
How this entry was checked and changed
Why this level

rubric's SR/MA with mostly small trials (2058 participants over 67 studies is about 31 per study, our arithmetic; the abstract gives no trial sizes); modelled dose-response, so slopes are not tested doses, and a set-counting rule chosen from the data. Only the abstract was read, so risk of bias, heterogeneity and the fitted curves could not be checked