Loading fundamentals

The question this branch answers

Most training advice is organized by variable. How long to rest, where to stop a set, how often to train a muscle. Each gets argued separately, each has its own camps, and each has produced its own body of research.

Read that research together and something odd emerges. Those three keep resolving into the same underlying quantity, and the recommendations that circulate for each of them were built on mechanisms that have since been tested and did not hold up.

The question this branch answers is not what the right numbers are. It is what those numbers are adjusting.

What is settled

Volume drives hypertrophy, and it has a graded dose-response. More weekly sets produce more growth, with diminishing returns as the count rises (Schoenfeld et al., 2017). This is the most robust finding in the branch and the one rest, set endpoint and frequency keep pointing back at.

Rest, set endpoint and frequency turn out to be levers on volume, not independent inputs. Rest duration changes how many repetitions survive to your later sets. Set endpoint changes how much capacity survives to the sets after it. Frequency changes how much weekly volume you can deliver without any single session running past the point where sets stop counting. The quantity all three adjust is how much non-degraded work you accumulate in a week.

Three widely repeated recommendations rest on mechanisms that were subsequently tested and failed. Short rest for hypertrophy came from the hormone hypothesis, and neither load nor systemic hormone elevations were found to determine training-mediated hypertrophy (Morton et al., 2016). Higher frequency came from the protein synthesis time course, which predicts a volume-equated difference that is not present (Schoenfeld et al., 2019). Training to failure came from motor unit recruitment reasoning, which predicts an advantage that did not appear in the most experienced participants tested (Refalo et al., 2024). In each case the mechanism is real at the cellular level and does not predict the outcome it was invoked to explain.

And in each case the superseded recommendation is still the one in circulation. The 30-to-90-second rest prescription, the "at least twice a week" rule, and the instruction to take every set to failure are all still standard advice. The first two have been contradicted by the same research groups that produced them (Singer et al., 2024; Schoenfeld et al., 2019).

Fatigue within a session is measurable and larger than most people assume. Taking sets to failure roughly doubles the capacity lost by the end of a session compared with stopping one repetition short (Refalo et al., 2023). This is the mechanism tying the branch together, and the reason rest, set endpoint and frequency cannot be adjusted independently of each other.

Where it splits

Rest between sets. Longer rest produces slightly more hypertrophy, and the difference disappears when volume load is equated. The plateau sits around ninety seconds, which is where the repetition cost stops accumulating. The practical consequence is that cutting rest is the worst available way to save time. (→ Loading fundamentals 01)

Reps in reserve. Failure costs roughly double the within-session capacity and did not produce more hypertrophy than one to two repetitions in reserve in participants averaging seven years of training. The reading error on the dial is around one repetition, biased toward underprediction, which argues in the same direction. (→ Loading fundamentals 02)

Training frequency. Frequency effects vanish under volume equating, for both hypertrophy and strength (Grgic et al., 2018). What survives is the interaction with a per-session ceiling: frequency stops being a preference once weekly volume exceeds what one session can productively hold. (→ Loading fundamentals 03)

Load and repetition range. Hypertrophy appears load independent and strength does not. The load-independence result was generated exclusively in trials taking every set to failure, which is the condition this branch argues against, so the two findings have to be reconciled rather than stacked. (→ Loading fundamentals 04)

Progressive overload. Three different claims sharing a name. Load progression and repetition progression came out comparable in the one randomized test, which is what you would expect if both are proxies for relative effort. Whether progression is strictly necessary has been tested once, in beginners over eight weeks, and not at all in trained lifters. (→ Loading fundamentals 05)

Exercise selection. Muscles grow unevenly, and which movement you pick changes where within a muscle the growth lands. (→ Loading fundamentals 06)

Weekly volume. The variable most of the branch resolves into. A graded dose-response with diminishing returns, running much further for size than for strength, and unanswerable as a number until you fix how indirect sets are counted. (→ Loading fundamentals 07)

Warm-up. One trial found the recommended specific warm-up made performance worse on isolation exercises, and another found it made no difference on bench and leg press. Ascending sets prepare coordination and calibration rather than tissue, so they are worth doing where those are hard and a cost where they are not. (→ Loading fundamentals 08)

Movement velocity and intent. The entry for velocity-based training and the dynamic effort approach. Two variables, separately isolated: how hard you try to accelerate changed adaptation in trials where the load was identical, and how far velocity decays within a set splits the outcomes. (→ Loading fundamentals 09)

Time off and coming back. Three to six months of cessation did not produce a detectable loss of muscle size in the population studied, and full strength returned in under eight weeks of retraining. (→ Loading fundamentals 10)

Estimated and tested maxes. A tested max is a reliable measurement. The case for estimating it is that a tested number describes one day and has not been shown to produce better strength gains than loading by effort. (→ Loading fundamentals 11)

Exercise order. The first exercise gets the most repetitions and the largest strength gains, and order has not been shown to change how much muscle you grow. (→ Loading fundamentals 12)

Adaptation weeks and deloads. A way in and a way out: early strength is largely learned, and some fatigue outlasts a rest day or two. Neither week has been shown to add muscle by itself, and a week of complete rest mid-block cost trained lifters some strength, so the deload keeps training. (→ Loading fundamentals 13)

How rest, set endpoint and frequency interact

If a single session has a ceiling on productive sets, then your weekly volume is bounded by frequency multiplied by that ceiling. And the ceiling itself is not fixed. It moves with the other two variables. Short rest lowers it, because repetitions collapse faster across sets. Training to failure lowers it, because capacity collapses faster across sets. A lifter with sixty-second rest periods taking everything to failure has a much lower per-session ceiling than one resting two minutes and stopping short, and will hit it sooner.

That produces a specific failure mode, and it is common. Someone tries to raise weekly volume by adding sets to an already long session, in a program with short rest and failure endpoints. The added sets land past the ceiling, do little, and add fatigue that degrades the following session. The volume went up on paper and not in effect.

The reverse ordering is the one the evidence supports. Set weekly volume, protect the ceiling by resting adequately and stopping short of failure, then use frequency to distribute whatever the sessions cannot hold.

This is the reasoning behind how OMNID programs are structured: rest long enough for the next set to count, most working sets stopped short of failure, and weekly volume spread across several sessions instead of concentrated into one. Those are not three separate decisions. They are one decision applied three times.

What this branch does not cover

Range of motion and tempo. Each will get its own entry. The range of motion evidence is currently moving faster than we can verify it, with the best-powered results sitting in preprint, so we are holding that one.

A caveat that applies to all of it. Almost every trial cited here measured muscle size at one or two sites. Growth within a muscle is uneven, so those numbers describe the sites scanned rather than the whole muscle. The precision of every effect size in this branch is worse than the decimal places suggest. (→ Loading fundamentals 06)

A concentration worth knowing about. Six of the seven sources underpinning this branch share an overlapping author network centered on one laboratory and its collaborators. That network produced most of the volume-equating work these conclusions rest on, and it has publicly reversed its own earlier positions more than once, which counts in its favor. It also means the independent replication that would move these findings from well-supported to settled has largely not happened. Read the effect sizes here as one research program's best current estimate rather than as a field consensus.

Whether you keep training at all. The variables here produce effects between a tenth and a third of a standard deviation, and whether you keep training decides whether any of them arrive. (→ Adherence 01: The first month)

Individual variation. Everything here is a group-level finding, and nobody has measured yours. What that warning means, and why your own twelve-week result is noisier than it looks, is covered separately. (→ Individual variation 01: Are you a non-responder?)

How to use this branch

In one sentence: train with enough rest that your later sets still count, stop most sets a repetition or two short, and add frequency rather than session length when you want more volume.