Weekly set counts are a speedometer, and a training year needs a map.

Three blocks, more sets every time, and a worse month eleven

Picture a lifter who did everything the volume literature seems to ask for. Three hypertrophy blocks across a year. Each one started a little higher on weekly hard sets than the last. The log is clean, RPE is honest, nothing got skipped.

And at month eleven the working sets feel heavier than they did in March. Knees are cranky. The top set on squat hasn't moved in five weeks, and the lifter can't say why, because every weekly number they checked said the plan was working.

That lifter is hypothetical. The shape isn't. Anyone who's run back-to-back blocks on a progressive-volume template will recognize it.

The weekly view never flagged a problem. It couldn't. Each week got compared to the week before it, and the week before it was always a little lower.

The speedometer problem

Weekly set counts are a speedometer, not a map. They report how fast you're going right now. They say nothing about the route, where you've been, or which roads you never took.

That's the speedometer problem. The unit most lifters use to manage volume only has meaning inside a week, yet the adaptations that matter (a bigger quad, a stronger bench, tendons that tolerate the load) accrue over months and years.

So the questions worth asking at month eleven are annual ones. How much hard work did each muscle group get this year? Where did it come from? Which weeks built something, and which were just accumulated fatigue with a spreadsheet attached?

A weekly number can't answer any of them. Add up fifty-two of them and you still can't, unless the log kept a consistent unit all year.

Weekly set counts report how fast you're going. They say nothing about the route.

Why sets times reps times load became the default

Tonnage became the default for one reason: it's the easiest number a log can compute. Every set already carries a weight and a rep count. Multiply, sum, chart. Hard-set counts won the next round for a similar reason. Count the sets taken close to failure and you're done.

Neither was picked because it describes a training year well. Both describe a database row well.

That gap is small inside one block. The lifter is running roughly the same exercises, rep ranges and proximity to failure for weeks at a time, so almost any unit tracks the trend. Across a year, everything moves. Exercises rotate. Rep targets swing between blocks. A peaking phase cuts volume while intensity climbs. The unit you chose in January is quietly measuring something different by October, and the chart draws it as one continuous line anyway.

What the dose-response research covers, and what it doesn't

The best-known evidence for "more sets, more growth" is the 2017 meta-analysis by Brad Schoenfeld, Dan Ogborn and James Krieger in the Journal of Sports Sciences, which found a graded relationship between weekly sets per muscle group and hypertrophy. Later work, including the dose-response meta-regression from Pelland and colleagues Pelland et al.'s meta-regression on resistance training volume/frequency dose-response was published in 2025, first appearing as a 2024 preprint ([source](https://link.springer.com/article/10.1007/s40279-025-02344-w)), points to diminishing returns as weekly sets climb.

That evidence has a timescale. The pooled trials are built from individual studies that each run for a matter of weeks or months, not years. They describe how a block responds to a dose. They don't describe three progressively larger doses stacked back to back.

Then there's counting. Does a set of bench press count as a full set for triceps, half a set, or nothing? Researchers have started testing fractional counting Pelland et al.'s resistance training dose-response meta-analysis found distinguishing direct vs indirect (fractional) sets is essential for predicting training adaptations. ([source](https://sportrxiv.org/index.php/server/preprint/view/460)), and whichever rule you pick moves the annual total for every muscle that mostly gets indirect work.

The muscle-group gap hiding inside a year that looks fine

Back to the hypothetical lifter. Say the year is three blocks of roughly four months each. Block one is a squat-and-bench base. Block two is an upper-lower hypertrophy split. Block three is a bench specialization with lower-body work held at maintenance.

Total weekly sets look healthy. Every block ran a higher weekly total than the one before, so any chart of sets per week slopes up and to the right.

Now split the year by muscle group. Hamstrings got real direct work only in block two, because block one leaned on squats (which don't do much for hamstrings) and block three parked the lower body. Upper back rode along on fractional credit from a few rows. Quads look fine on paper, but most of those sets were heavy triples at RPE 7, a long way from failure.

So the aggregate says progression. The per-muscle picture says hamstrings had a four-month year. The lifter at month eleven has plenty of volume in the log. It's sitting in the wrong places, and nothing in the weekly view ever showed it.

What a year-level volume picture requires from a log

Four things, and most logging interfaces are built around only one of them: the week.

Per-muscle attribution comes first. Every exercise maps to the muscles it trains, with a declared rule for indirect work, applied the same way all year. Second, an effort field that survives. RPE or RIR on working sets, so a year-level tally can separate hard sets from sets that just happened.

Third, block boundaries. The log has to know where one phase ended, or twelve months collapse into one undifferentiated line.

Fourth, continuity. The whole year in one place, including the months you logged in a spreadsheet or a different app before you switched.

Miss any of the four and the annual picture is guesswork. The position that falls out of all this: plan volume weekly, audit it annually, and treat the audit as the thing that decides next year's first block.

The year has to be in one log before anyone can read it

Continuity is the dull requirement, and it's the one that breaks most annual pictures. Lifters switch apps. They run a block from a spreadsheet, then drift back. The year ends up in three places and no single chart can see it.

Platepusher starts there. A CSV from Strong or Hevy imports every lift, every set, every date as native data, so the months before the switch sit in the same record as the months after. Plateau detection flags the week a top set stops moving, which is the moment the lifter in this piece needed: five stalled weeks in, with a whole year of history ready to open.

The instrument keeps the record. Reading the year is still the lifter's call.