Load is a proxy for muscle: hypertrophy hides in your RPE column.

The bench went up 10 kg. Did the chest?

Adding 10 kg to your bench over a training year proves you got better at benching. Whether your chest got any bigger is a separate question, and the bar can't answer it.

Most lifters running a hypertrophy block know this in theory. Then they open their tracker, see an estimated 1RM line sloping up, and quietly read it as a muscle chart. It feels like the same thing. Some months it is.

Other months the line climbs because your setup got tighter, your arch got bigger, or the last rep started bouncing off your sternum. And some months it sits dead flat while you're adding real tissue at the same load with more left in the tank. Both cases produce the same chart.

That's the problem.

The proxy gap

Call it the proxy gap. Load progression stands in for hypertrophy, and strength-first trackers quietly treat the stand-in as the measurement.

Strength has a clean number: weight on the bar, reps completed, an estimated max you can plot week over week. Hypertrophy doesn't have one. Muscle size changes slowly. Research labs measure it with ultrasound or MRI, and nothing on a commercial gym floor reads it directly between sets.

So trackers plot what they can count. That part is reasonable. The error comes later, when a lifter whose goal is size reads a strength chart as a verdict on size and rewrites their program because of it.

Strength has a number. Hypertrophy has a trail of evidence, and your log is where the trail lives.

Why "just add weight" is only half the answer

The standard advice says progressive overload drives growth, so if the weight is going up, you're growing. That's half right. Mechanical tension drives hypertrophy, and adding load is one way to raise it. But load is just one input to how hard a set actually was. Reps, proximity to failure, range of motion, tempo and accumulated fatigue all change the effort of a set, and a load-only log throws most of that away.

The usual workarounds don't close the gap either. Scale weight lumps muscle, water, glycogen and fat into one number. A pump is blood flow from the same day and tells you nothing about next month. Progress photos and tape measurements are worth taking now and then, but they're noisy from one session to the next, and lighting does half the work in a photo.

Every one of those is a snapshot. Hypertrophy shows up as a trend over months. The only record that's there for every set you do is the log.

What the research says when strength and size move apart

Strength and size coming apart isn't just gym folklore. It shows up in controlled training studies.

In a 2012 study in the Journal of Applied Physiology, Mitchell and colleagues had young men train to failure at either 30% or 80% of 1RM. Muscle growth was similar across the two conditions, but 1RM gains favored the heavy condition. Schoenfeld and colleagues found the same pattern in trained men in 2015 (Journal of Strength and Conditioning Research). They compared sets of roughly 25 to 35 reps against sets of 8 to 12 and saw similar hypertrophy, with larger strength gains in the heavier group.

Taken together, the two studies show that a lifter's max can move independently of their muscle growth, depending on how they train. Buckner, Loenneke and colleagues went further in a 2016 paper in Muscle & Nerve, arguing that the field has assumed a tighter causal link between growth and strength than the evidence supports.

There's a second detail in those designs. The light-load groups only matched the heavy groups on growth because their sets went to failure. Effort carried the result, and effort is exactly the variable a load-only log never stores.

Two logs, same lifter

Take a lifter running a 12-week hypertrophy block, with incline dumbbell press as a main movement. In week 1 they log the 32s for three sets of 10. In week 8 they log the 32s for three sets of 10. A load-only log shows eight weeks of nothing, and if the tracker has a plateau flag, it fires.

Now add the RPE column. In week 1 the top set was an RPE 9 and the third set fell to 8 reps. By week 8 all three sets hit 10 at RPE 7 to 7.5, with a note that the last reps stopped grinding. Same load, same reps, about two more reps in reserve. Capacity went up at the same tension, and the log caught it.

Flip it around. A different eight weeks shows the load climbing from the 32s to the 36s. That looks great until you read the notes: "touched shoulder, lost the stretch" and "cut ROM last two reps," with RPE 9.5 logged every week. The chart says progress. The log says the set got shorter as it got heavier.

You can't see either story on a load chart. Both are plain in a log that treats RPE and set notes as data.

How do you know if you're building muscle?

If the proxy gap is real, the question changes. You stop asking whether a number went up this week. You start asking whether your capacity at a given effort is rising over months.

In a log, that shows up three ways: more reps at the same load and RPE, the same reps at a lower RPE, or more load at the same RPE with notes confirming the range of motion held. Any of those, holding across several movements over 6 to 12 months, is as close to a hypertrophy readout as a gym record gets. One great session tells you very little. Twelve weeks of RPE slowly drifting down for the same work tells you a lot.

It also changes what counts as a plateau. For a size-focused lifter, flat load with falling RPE is progress that looks like a plateau. Rising load with rising RPE and shrinking range is a stall that looks like progress. A tracker that can't tell those apart gets the verdict wrong both ways.

Where Platepusher sits in this

Platepusher starts from the idea that a lifter's own history is the most useful dataset they'll ever have. For hypertrophy, that only holds if the history keeps more than the load.

RPE sits in the set row next to weight and reps, and it's logged as fast as the numbers Per-set notes are a current feature: a note can be attached to any individual set, and it carries forward to the same set in the next session of that plan workout. Per-set RPE is not user-facing in the current build: the RPE and %TM editing UI was removed, and RPE survives only as dormant stored data on plans and sets. Plateau detection does not read RPE. It works from logged load and reps (with heart-rate fatigue signals when an Apple Watch is linked), and it excludes sessions flagged as deloads.. The plateau flag describes what the math sees across weeks. It doesn't hand you a program or pose as a coach. And if your last few years live in another tracker, the CSV import brings every set across as native data, so the 6-to-12-month trend starts with the years you've already logged.

You won't get a body-composition guess or a muscle-gain score. The app shows the trail, and reading it stays your call.