Guide
How an estimated 1RM is calculated — and when it lies
Almost nobody tests a true one-rep max, and nobody tests one often enough to track progress with it. An estimate turns every working set into a comparable number instead — which is useful right up until you ask it a question it cannot answer.
The formulas
The two you will meet everywhere:
- Epley:
1RM = weight × (1 + reps / 30) - Brzycki:
1RM = weight / (1.0278 − 0.0278 × reps)
Both fit a straight line to a relationship that is not straight. Near a true max they agree closely: 100 kg for 5 gives 116.7 kg by Epley and 112.5 kg by Brzycki — a 4% spread, small enough not to matter if you stay with one formula.
Far from a max they fall apart, in opposite directions. Take 60 kg for 20 reps: Epley says 100 kg, Brzycki says 127 kg. Neither is a measurement. High-rep sets are limited by breathing, grip and pain tolerance long before they are limited by the same thing that caps a heavy single.
What a rep cap fixes
Left alone, this produces a specific absurdity: a 20-rep back-off set out-ranking the heavy triple that preceded it. By Epley, 60 kg × 20 reads as 100 kg while 100 kg × 3 reads as 110 kg — close enough that one bad day flips the order, and your "best set of the session" becomes the easiest one.
Hevy Analyzer uses Epley with the rep count capped at 12. The same back-off set is then scored as 60 × (1 + 12/30) = 84 kg, and the heavy triple keeps the session. The cap does not make high-rep sets estimate correctly — nothing does — it stops them from overwriting the sets that carry the signal.
Bodyweight work has no external load, so the estimate is zero and no trend is reported at all. A profile built on pull-ups and dips will show plenty of sets and no 1RM line, which is honest: there is nothing in the log to estimate from.
When the number lies
- High reps. Past about 12 the estimate is a genre of fiction. This is the big one.
- Different proximity to failure. Five reps with three in reserve and five reps to failure are not the same set, and the formula cannot see the difference.
- Machines and cables. Leverage, pulley ratios and stack labels vary between gyms, so the load on the plate is not the load on you.
- Fatigue within a session. The same set moves down the more work precedes it. Compare like for like — same slot in the session, ideally.
- Technique drift. A shortened range of motion inflates the estimate exactly when you are least likely to notice.
Using it properly
Treat the estimate as an index, not a prediction. Two rules make it behave: compare the same exercise, and compare within a similar rep range. Then the absolute number can be wrong by 10% and still tell you the truth about the direction, because the error is roughly constant and cancels out.
That direction is the whole point, and it is what plateau detection reads: session estimates in sequence, recent half against the half before it, with guards so noise does not get promoted to a conclusion.