Guide
How to tell if a lift has actually plateaued
A bad session is not a plateau. Neither are two. Bar speed, sleep, the order you ran the exercises in and whether you ate lunch move a top set around by more than most real stalls do — which is why people announce plateaus that resolve themselves the following week, and miss the ones that have been running for two months.
The problem with judging it by feel
You remember your best session and your worst one. The forty in between are the trend, and they are exactly the ones memory drops. Worse, a plateau and a hard week look identical from inside a single session — the difference only exists across time, which means it can only be read from a log.
A method that survives noise
Pick one exercise. For each session, take the best set and convert it to a single number you can compare across different weights and rep counts — an estimated one-rep max. The Epley formula is the usual one:
estimated 1RM = weight × (1 + reps / 30)
Then compare your last three sessions against the three before them. If the recent average is meaningfully higher, the lift is moving. Meaningfully lower, it is going backwards. Inside a couple of percent either way, it is flat.
Past roughly 12 reps the formula drifts high, so a 20-rep back-off set can out-rank a heavy triple if you let it. Cap the rep count you feed it, or only ever compare sets in a similar rep range.
Three guards worth keeping
- At least six sessions. Below that you are reading noise with extra steps.
- At least two weeks of spread. Six sessions crammed into eight days is enough data points and not enough time — nothing has had a chance to adapt.
- A dead band. Anything inside about ±2% is flat, not a change. Without this every lift is either "progressing" or "declining" every single week, which tells you nothing.
If a lift fails any of those, the honest answer is not enough evidence — not "plateau". An exercise you train once a fortnight will trip a naive plateau detector constantly, and the detector will be wrong every time.
Hevy Analyzer applies exactly this to any public Hevy profile: best estimated 1RM per session, last three against the previous three, six-session minimum, 14-day minimum spread, ±2% dead band. Each exercise gets a Progressing / Plateau / Declining label with the percentage, the session count and the day range behind it, so you can check the claim instead of trusting it. Bodyweight movements produce no load estimate, so they report no trend at all rather than a fake one.
So it is a plateau. Now what?
Ranked by how often each turns out to be the real cause:
- Recovery, not the program. Sleep and calories stall lifts faster than any set scheme fixes them.
- Attendance. Check the actual session count over the last eight weeks against what the program asks for. This is the one people are most surprised by.
- Volume that never changed. The same three sets that got you here are not obliged to take you further; a few more weekly sets for that muscle is the usual next step — see sets per muscle per week.
- Fatigue you have been carrying. A lighter week is not lost progress; for a stalled lift it is frequently the thing that unsticks it.
- The exercise, not you. Long-run stalls on one movement often break when you rotate to a close variation and come back in six weeks.
Change one of these at a time. Change three and you will not know which one worked, and you will be back here in three months guessing again.