How much work
actually does anything

Ten sets a week for chest. Twenty for legs. You have read the numbers and so has your client, usually on a forum, usually without the sentence that makes them mean anything.

Before any of those numbers can be useful, you have to answer a smaller question: when your client benches, how many chest sets is that, and how many triceps sets? Most volume advice never says, and the answer changes the recommendation by a factor of two.

Direct work and indirect work

A bench press trains the chest directly. It also trains the triceps and the front delts, but not in the same way and not to the same degree. A row trains the lats directly and the biceps and rear delts along the way.

So there are three ways to count a week, and they give three different answers for the same training:

This is not a semantic argument. Five sets of bench and five of dips is ten triceps sets by the first method, seven and a half by the second, and five by the third. If your target is ten, you have either hit it, missed it by a quarter, or missed it by half, and nothing about the training changed.

What the evidence actually says

This was tested directly rather than argued about. Pelland and colleagues, in Sports Medicine in 2025, pooled 67 studies and 2,058 participants and ran the same models three times, once under each counting method, to see which one best predicted what actually happened to people.

For weekly volume, the fractional method won. Counting an indirect set as half a set had the strongest relative evidence and was used for their primary models. Their conclusion was blunter than the result: distinguishing direct work from indirect work appears to be essential for predicting adaptation at all.

The argument was never ten sets against twenty. It was that ten means two different things depending on who is counting.

Where it stops holding, which matters more if you coach strength

The companion analysis found something different, and most write-ups of this research leave it out. Remmert and colleagues ran the same comparison on volume per session rather than per week:

Read plainly: if the goal is a bigger muscle, the accessory work counts for something. If the goal is a heavier lift, the evidence per session says the sets that move that lift are the ones doing the work, and the indirect sets are worth less than half.

That is a real split rather than a rounding difference, and it is the one most volume guidance skips, because "count indirect work as a half" is a tidier sentence than "count it as a half, unless you are chasing a number on the bar, in which case possibly do not count it at all".

The honest ceiling

Both papers found the same shape: more work does more, and the return per set falls off. The per-session analysis put the useful ceiling at around eleven fractional sets in one session, with clear diminishing returns beyond it. Not a cliff, not a limit, just the point where each additional set buys noticeably less than the one before.

Why the counting method is a coaching decision, not a technicality

Two clients, identical programmes, one chasing a bigger back and one chasing a heavier deadlift. The same sheet of paper is adequate volume for one and possibly not for the other, and the only thing that changed is which counting method the evidence supports for their goal.

It also decides what you do when a client says they want more work. Adding accessories raises the count under fractional counting and barely moves it under direct counting, so the same addition is either a meaningful increase or almost nothing, depending on what they came for.

What this looks like in Trainersight

The app counts fractionally: a direct set counts one, an assisting set counts a half, and the per-muscle share on an athlete's Trends screen falls out of that rather than being typed in by anybody. It is the same rule Renaissance Periodization publishes, and it is the weekly model the evidence above supports.

It is a weekly model shown to strength athletes, which is exactly the case the split above is about. Worth saying rather than leaving in the code.

The short version

• A set is not a set. Direct work and indirect work are different things and the difference is measurable.
• For weekly volume, count an indirect set as half. That is the model with the best evidence across 67 studies.
• For strength, per session, the evidence favours counting only the direct work.
• Returns diminish past roughly eleven fractional sets in a session.
• Any volume number without a counting method attached is missing the half that makes it actionable.