How do I count training volume for hypertrophy?

Count hard sets per muscle group per week: sets you finish within roughly five reps of failure. Greg Nuckols and Stronger By Science argue this is the predictive unit, and Schoenfeld's dose-response work counted weekly sets per muscle rather than tonnage. Tag each working set with RIR or RPE so the count is recoverable later.

Hard sets vs total sets for muscle growth, what's the difference?

Total sets treats an easy back-off set and a set taken to the brink as identical. Hard sets only counts the ones that got close enough to failure to drive growth, roughly within five reps. That distinction is why total set count stops meaning anything when you move from high-rep volume work to heavy low-rep work.

Why did my volume go up but I stopped making gains?

Usually because the volume that went up was tonnage or raw set count, neither of which tracks how hard the sets were. You can add tonnage by adding easy back-off reps and grow your graph without adding a single hard set. Track hard sets per muscle instead and the number lines up with the stimulus that actually causes growth.

What counts as a hard set, and how close to failure does it have to be?

The working definition researchers tend to use is a set finished within about five reps of failure, the range where the stimulating reps sit. In practice that means logging the set at roughly RPE 8 or higher, or 3 RIR and below, and counting those toward your weekly total per muscle group.

Hard Sets Are the Only Volume Metric That Transfers Across Programs

Article · 4 min read

Hard sets are the only volume number that survives a program change.

Total sets and tonnage break every time you switch programs. Hard sets per muscle per week is the one volume number that stays comparable.

The number that lied to you

You finish the last set of a twelve-week hypertrophy block, 5x10 at 60 percent, and the session tonnage on your log reads north of four thousand kilos. Two weeks later you're deep in a strength block, 3x5 at 85 percent, and that same graph has fallen off a cliff. Less than half the tonnage. Two fewer sets. The line points down like you've been detraining.

You haven't. You've been getting stronger. The number just stopped measuring anything real. This is the quiet failure at the center of almost every tracker: the volume figure it hands you is bound to the program that produced it, and it silently invalidates itself the moment you change loading schemes.

The only portable volume unit

Volume is real and it drives growth. The way most apps count it is program-specific, and that's why the number breaks every time you change rep ranges. The one measure that stays comparable from a 5x10 block to a 3x5 block to an RPE-capped block is hard sets per muscle group per week: sets you finish within roughly five reps of failure.

Call it the portable unit. It's the same dose whether you got there with tens at 60 percent or fives at 85. Everything else on your dashboard is a snapshot of one program, useful inside that block and meaningless the second you leave it.

Tonnage and set counts are program-specific by construction

Tonnage is sets times reps times load. Move to heavier weight and lower reps and the product drops even when the training gets harder. It measures mechanical work. Growth stimulus is a different thing, and the two diverge the moment your rep ranges change.

Raw set count has the opposite flaw. Five sets of ten easy back-off reps and five sets taken to the brink collapse into the same integer, three-plus reps of difference erased. Neither number knows how close you got to failure. And proximity to failure is exactly what the hypertrophy literature keeps circling back to.

Neither tonnage nor a raw set count knows how close you got to failure, and proximity is what the muscle actually responds to.

What Nuckols and the hypertrophy literature actually count

Greg Nuckols and the Stronger By Science team made the case cleanly in their reframing of training volume: the predictive unit for hypertrophy isn't tonnage and it isn't raw set count, it's the number of hard sets per muscle group. Brad Schoenfeld's dose-response meta-analysis, the one everyone reaches for to argue 'more volume, more growth,' operationalized volume as weekly sets per muscle, not as tonnage.

When researchers count a set as 'hard,' they typically mean it finished within about five reps of failure, the zone where the stimulating reps live. Stack those two facts together and the conclusion is hard to avoid: the volume figure with the most research behind it is one almost no tracker logs by default.

Illustrative, one lift, a 140 kg squat 1RM. Tonnage swings more than 2x and total sets vary block to block; hard sets is the figure you'd actually equate across all three.
Loading block	Session tonnage	Total sets	Hard sets (~5 from failure)
5x10 @ 60% (84 kg)	4,200 kg	5	2
3x5 @ 85% (119 kg)	1,785 kg	3	3
RPE 8 triple + 3x5 backoffs	~2,050 kg	4	4

What the log entry has to look like

The fix is a notation change you can make on your next set. Capture RIR or RPE on the sets that matter, in the two seconds after you rack the bar, so the hard-set count is recoverable later instead of gone.

Without that field, three logged sets of 120x5 are just three sets. You can't tell later whether they were grinders or warm-ups wearing working weight. With it, the weekly rollup can count how many landed inside the effective range, and that count reads the same whether last month was a volume block or a peaking block.

Same three sets, two notations. Only the right column survives a move to a different rep range.
What you capture	Without RIR	With RIR
The set log	Squat 120x5, 120x5, 120x5	Squat 120x5 @8, x5 @9, x5 @9.5
What it tells you later	3 sets, load, reps	3 hard sets, all within 2 of failure
Comparable across a program switch	No	Yes

Your longitudinal graph starts meaning something

Change the unit and the multi-year log you've been building stops resetting its meaning at every mesocycle boundary. Twelve hard sets for quads in a hypertrophy block and twelve in a strength block are genuinely the same dose, so a drop from twelve to eight is a real signal you can act on, not an artifact of switching rep schemes.

That's the line between a training history you can actually run experiments on and a pile of numbers that go blank every eight weeks. For a lifter three or four years in, whose whole reason to log is comparing this cycle against the last ten, that comparability is most of the value.

Logging the unit, not just the sets

Most trackers count sets and multiply out tonnage because those numbers come for free from weight and reps. Proximity to failure doesn't. It has to be captured at the set or it's lost, and that's the field almost everyone drops.

Platepusher keeps RIR and RPE as first-class fields on the set, and the weekly view rolls them up as hard sets per muscle group. So the number that transfers is the number on the chart. The instrument records what you did near failure. It doesn't tell you to train harder or push a plan on you. It shows you where the stimulus actually landed, block over block.

What we're still working out

The open question is attribution. A hard set of Romanian deadlifts hits hamstrings, glutes, and spinal erectors, and the field hasn't fully settled how to credit a compound set across the muscles it works. Count it whole for each? Fractionally for the secondary ones?

The practical default most volume-counting frameworks use is a full set for the muscles heavily involved and a partial credit for secondary ones, and that's the convention worth watching as the research tightens. When it does, the rollup should follow it rather than lock in a rule the evidence hasn't earned.

Log RIR on the set that matters and let hard sets per muscle roll up on their own. Get Platepusher and keep a volume number that survives your next program change.

Serious lifters three and four years in have logs full of tonnage graphs that reset their meaning at every program switch. Counting hard sets per muscle is the fix, and it's a change you can make on your next session's very first set.

Get Platepusher