Your training log is a lagging indicator of the work you did weeks ago.

Week 6, same number, different set

Week 6 of a squat block. Since week 3 the program has called for 315 for five on the top set, and that's exactly what the log shows: 315x5, 315x5, 315x5, 315x5. A month of flat line. The lifter scrolls back, sees the same number four times, and starts reading program reviews on the bench between sets.

But tonight the bar moved differently. Faster out of the hole. The fifth rep didn't grind. In week 3 that set was an honest RPE 8, two reps left and not a hair more. Tonight it was closer to a 6.5, and the log says so, in the RPE column sitting right beside the load.

The load column is the one that gets read. And it's reporting on training that happened weeks ago.

The Lag Rule

Call it the Lag Rule. Whatever week 6 of a block builds doesn't show up in the log during week 6. It surfaces later, once that block's fatigue clears and the structural work consolidates. Usually that's the week after a deload, or the first test day of the next block.

So a log read as a live scoreboard is biased, and biased in one direction. It under-reports progress while the block runs and then seems to over-report it right after the block ends. Every mid-block read leans toward a false negative. That's the read that gets a lifter to drop a program that was working.

How long is the delay? Weeks. Not days, and not a clean number you can circle on a calendar. It shifts by lift, by lifter, and by how much fatigue the block deliberately piles on.

The load PR is the receipt. The adaptation was paid for weeks earlier.

Where the scoreboard habit comes from

Most lifters learned to read a log during their novice phase, and back then the scoreboard read was correct. Add five pounds, hit it, add five more. Early strength gains are mostly neural: better coordination, better motor unit recruitment, a groove that cleans up session by session. That kind of improvement shows up almost immediately, so the log tracked the lifter in near real time.

The habit outlasts the phase. By year three the lifter is running 5/3/1, a block-periodized peak, or an RP-style hypertrophy mesocycle, and those programs are built to accumulate fatigue on purpose. During accumulation the top set is supposed to look stuck. The program's author knows that. The lifter reading the log on a Tuesday night often doesn't.

This also needs separating from noise. Bad sleep, a long workday, a warm gym: those scatter the numbers both ways, and across enough weeks they mostly cancel. Lag doesn't cancel. It points the same way every week of the block, which is why reading the log too early produces the same wrong conclusion every time.

How long does it take for strength gains to show up?

The research gives each signal its own clock.

Moritani and deVries (1979) tracked the time course of strength gain and found the early weeks dominated by neural factors, with hypertrophy contributing more as training went on. Seynnes and colleagues (2007) detected measurable muscle growth within the first few weeks of high-intensity training, earlier than had been assumed. Then Damas and colleagues (2016) complicated that early number: much of the early increase in muscle cross-sectional area coincided with edema-like swelling from muscle damage. So the early tape-measure reading isn't clean either.

The piece that explains the lag in the log itself is older. Banister's fitness-fatigue model (Banister et al., 1975) treats performance on any given day as fitness minus fatigue. Fatigue builds fast and clears fast. Fitness builds slowly and sticks. Pile on volume and fatigue masks the fitness underneath; pull volume back and the fitness shows. Tapers and deloads run on that premise.

Which leaves one signal readable in real time: effort at a fixed load. The reps-in-reserve RPE scale described by Zourdos and colleagues (2016) gives lifters a consistent way to record it.

A working block, read from inside the log

Back to the squat block, as an illustration. The program holds the top set at 315x5 from week 3 through week 8. The load column never moves. The RPE column reads 8 in week 5, 7.5 in week 6, 7 in week 7, and 6 in week 8.

That's the whole check. If logged RPE on a given weight drops from an 8 to a 6 across weeks 5 to 8, adaptation happened, whether or not the load has moved yet. And it happened while fatigue was still climbing, which makes the drop more telling, because the lifter got stronger faster than the block could tire them out.

Week 9 is a deload. Week 10 opens the next block heavier, and the set that would've looked like a PR jump in a scoreboard read is just the receipt arriving.

A real stall looks different, and the difference is easy to see once both columns get read. Load flat. RPE flat or creeping up across the same weeks, with sleep, depth, and bar held roughly constant. That pattern deserves a response. The first one deserves patience.

What changes when the log lags

Judge a block after its deload, not during it. A mid-block switch needs a stall signature in both columns, not just a flat load line.

Read two columns every time. The load column confirms. The RPE column predicts. A lifter who only reads the first is always a few weeks behind their own training.

Credit the right block. The PR that lands the week after a deload gets talked about as if the deload produced it. Mostly, the block before did the work and the deload let it show. Get that backwards and the next program change goes after the wrong variable.

And leave room for the confirmation. A block with no deload and no test day never gives the lagging signal a chance to arrive, so the log stays flat by design and the lifter blames the program.

A log that keeps both columns in view

Most strength trackers lead with load and volume charts. The RPE a lifter types in tends to stay buried inside a set row, which is exactly how a working block ends up looking like a stall.

Platepusher's plateau flag runs math on the top-set trend across weeks instead of reacting to one flat session, and per-set RPE sits in the same record No. Platepusher does not currently offer per-set RPE logging or an RPE-at-fixed-load trend view. RPE and %-of-max entry in the plan editor and workout screen was deliberately removed in v1.0.6, and it is slated for a possible later release. The underlying data fields are kept in the model for future use, but no in-app view trends RPE at a fixed load.. That way a flat load line can be read against the effort that produced it.

The longer payoff is history. A lifter who brings years of logs over from another app already has a dataset of their own lag: every past block, every deload, every post-deload jump. How many weeks did the receipt take to arrive last time? The answer is already sitting in that record, and it's more useful than any population average (more on this in Your training history is the experiment).

What we're tracking next

Two open threads. First, whether the lag runs longer on lower-body lifts than on pressing, as many lifters report anecdotally Greater recovery times are needed for lower-body exercises (48–72 hours) than for upper-body exercises. ([source](https://pmc.ncbi.nlm.nih.gov/articles/PMC11057610/)). Second, whether RPE drift at a fixed load reliably comes before the post-deload PR in multi-year logs, or just usually does.

The stance doesn't depend on either answer. A flat top set in week 6 tells you where you were a few weeks ago. Read the RPE column for where you are now.