Epley vs Brzycki: Which 1RM Formula Should You Trust?

Almost every one rep max calculator on the internet is built on one of two formulas: Epley (1985) or Brzycki (1993). They agree closely at low reps — which is why both survived — and drift apart as reps climb. Here they are side by side, with the actual numbers, so you can see exactly where each one earns or loses your trust.

The two formulas

Formula	Equation	Idea behind it
Epley	1RM = w × (1 + reps ÷ 30)	Each rep adds a flat 1/30th (~3.3%) of the weight lifted
Brzycki	1RM = w × 36 ÷ (37 − reps)	Each rep is a bigger slice of a shrinking remainder — the cost per rep grows

Both are on our one rep max calculator, alongside Lombardi, Mayhew, O'Conner and Wathan, so you can compare your own sets across all six.

The numbers: 100 kg lifted for 1 to 15 reps

Reps at 100 kg	Epley	Brzycki	Difference
1	103.3	100.0	3.3 kg
3	110.0	105.9	4.1 kg
5	116.7	112.5	4.2 kg
8	126.7	124.1	2.6 kg
10	133.3	133.3	0.0 kg
12	140.0	144.0	4.0 kg
15	150.0	163.6	13.6 kg

Three things stand out. First, at a single rep Brzycki returns exactly the weight lifted — mathematically tidy — while Epley adds 3.3%, which is arguably more realistic since almost anyone can single slightly more than a weight they took to a hard double's neighbourhood. Second, the two curves cross at exactly 10 reps. Third, beyond 10 reps Brzycki accelerates away — and that acceleration is a warning sign, not a feature.

Where each formula breaks down

Brzycki has a cliff edge. Its denominator is (37 − reps), so as reps approach 37 the estimate explodes towards infinity, and past 37 it goes negative. Long before that, in the 15–25 rep range, it produces numbers no coach would program from. Brzycki himself designed it for low-rep sets; treat anything above 10 reps as out of its warranty.

Epley is linear, which is its own lie. A flat 3.3% per rep assumes the 14th rep costs the same as the 4th. In reality each additional rep costs more than the last as fatigue compounds, which is why Epley starts underestimating endurance-strong lifters at moderate reps and why its predictions at 15+ reps are still only a ballpark.

The deeper point — explored in our article on how accurate 1RM calculators are — is that no formula fixes a bad input set. Both formulas were fitted on low-rep data from trained lifters, and both behave best exactly there.

So which one should you use?

• Sets of 1–6 reps: take your pick — or better, average them. The 4 kg gap at 5 reps is smaller than day-to-day strength variation.
• Sets of 7–10 reps: both are reasonable; the average is a sensible middle.
• Sets above 10 reps: trust neither fully, and distrust Brzycki more. If you must estimate from a high-rep set, Epley's answer is the more conservative one.
• Programming, not predicting: remember the estimate feeds your training max and percentages. A 3% formula disagreement changes a working set by a kilo or two — your programme will absorb it.

Our own approach on this site: show all six formulas and their average. When five of six cluster together, you have a number worth training from. When they scatter, the set itself was the problem — go and take a cleaner, heavier set instead.

Frequently asked questions