Dorian Yates logged effort, not volume. That changes what progress means.

The only set that counted

Temple Gym, Birmingham, mid-morning. The warm-up sets on the barbell row are done and forgotten. Then comes the one set Dorian Yates actually came to do: reps to failure, a forced rep dragged past it, then the bar down. That set is the single line in his diary his whole training week was built around. Everything before it was ramp. Everything after was recovery. Drop the same session into a modern tracker and it logs four working sets of rows, all equal, all summed into a tonnage figure. Yates logged one.

The effort unit

Call it the effort unit. Yates operationalized proximity-to-failure as the thing his log recorded: one all-out set per exercise, taken to failure and usually past it. The number of sets was small and fixed. What varied, and what he wrote down, was whether that one set beat the last one. Reps at the same load, or the same reps at more load, both carried to the same place. He was recording quality of effort per set decades before the RPE scale turned proximity-to-failure into a phrase lifters throw around on the platform.

What tonnage was never built to see

The orthodoxy of his era, and the default of nearly every app since, is tonnage. Sets times reps times load, summed across the week, plotted as a line that should climb. Volume-theater. It rewards the lifter who adds a fifth set of half-effort curls, because that set moves the total. It stays blind to the gap between a set left three reps short and a set taken to where the bar stops moving. A tonnage column cannot tell those apart. It was never built to.

What the record actually shows

Yates won six consecutive Mr. Olympia titles, 1992 through 1997, and retired as champion. He built that on a model he documented obsessively: warm-up sets to prime the joint, then one working set carried to failure and beyond with a forced rep or a slow negative. The approach drew on Mike Mentzer's Heavy Duty, which Yates has credited directly. The through-line in his own accounts of it is not how much moved in a session. It is whether the one set that counted got past where it sat last week.

The same session, two logs

Put one back session into a default set-logger and into Yates's diary, side by side, and the gap is not cosmetic. A default Strong or Hevy log records four rows of barbell rows, computes a tonnage figure, and treats the ramp sets and the death set as the same kind of data. Yates's page circled one line and wrote one question in the margin: did it beat last week. The forced rep lived in that margin. The app has no field for it. The app stored rows. Yates stored a verdict.

If effort is the unit

If effort is the unit, most of what a tracker proudly totals is noise. Two lifters, one logs 22 sets and one logs 6. One of them may have trained harder, and the tonnage column will tell you the opposite. For a beginner it rarely matters. Progress comes from almost anything. For a lifter five years in, whose top set climbs by a rep a month when the block is honest, the only signal worth capturing is the one Yates wrote down: how close to the wall was that set, and did it move. Advanced-lifter coaching calls progress at that stage watching paint dry. You cannot watch what your log does not record.

A ledger, or a dataset

This is the line Platepusher draws through its session logging. A set carries more than a weight and a rep count. It carries a record of how hard it actually was, so the working set reads differently from the warm-ups that fed it. A log that stores only load and reps counts what you lifted. A log that stores effort knows how hard it was, which is the thing a plateau flag or a progression call actually needs. Import a Strong or Hevy history and the old rows land as native data. From there, what gets recorded sits closer to what Yates kept than to what a tonnage total sees. The 2026 retrospectives on long-tenured lifters keep landing on the same metric, and the open question is how much of it a lifter will enter by hand, session after session, versus how much a log should carry for them. Yates entered all of it himself. Then again, he had six titles riding on the diary.

If your log can't capture how hard that set actually was, it's a ledger, not a dataset.