One-rep max calculator: estimate it from one hard set

A one-rep max (1RM) is the most weight you can lift once with good form. You rarely need to test it to use it: one hard set of a few reps carries most of the same information. Enter that set below.

Estimate your one-rep max

The calculator needs JavaScript, which looks to be switched off. The numbers here are the worked example, and the formulas below work by hand.

Training loads from your estimate

Loads for the same reps at each RPE

What the estimate expects for 5 reps at each effort level. RPE is 10 minus the reps left, the same scale as reps in reserve.

What a one-rep max is for, and why most people never test it

Strength programs often write loads as a percentage of a one-rep max: three sets of five at 80%, say. That needs a max, and testing one properly means a long warm-up, a few heavy singles, a spotter or safety bars, and a session spent on little else. The number also moves. A review of the reps-in-reserve scale notes that a one-rep max is not stable in new lifters, and that tiredness from earlier training can hold it down.¹ Unless a meet or a test day asks for the real number, an estimate from a recent hard set does the same job.

Where the formulas come from

The calculator uses two classic formulas and shows both:

• Epley: 1RM = weight × (1 + 0.0333 × reps). Each rep adds about a thirtieth of the weight.²
• Brzycki: 1RM = weight ÷ (1.0278 − 0.0278 × reps).³

Neither began as a study. Epley's comes from a 1985 poundage chart, Brzycki's from a 1993 journal article on strength testing. A 2014 paper that traced the usual rep tables to their sources calls the chart "not based on scientific research" and notes that the article doesn't say what its equation is based on.⁴ The studies below tested them later.

The two agree at 10 reps, where both put a set to failure at about 75% of a max. Elsewhere they split: by Epley a set of 5 is about 86% of a max, by Brzycki about 89%. Below 10 reps Epley gives the higher estimate; above 10, Brzycki does.

How accurate the estimates are

Close enough to train from, if the set is heavy. Across very different groups, the fewer reps in the set, the better the estimate.

• In about 220 men in seven groups, from untrained students to football players, six formulas tracked the tested bench press max closely (correlations of 0.82 to 0.98). From 10 reps or fewer, Brzycki's was the most accurate; above 10, all six ran significantly high or low. The authors suggest averaging two formulas may reduce the error,⁵ so the headline number here is the average.
• In 70 adults aged 18 to 69, 34 men and 36 women, a 5-rep max predicted the bench press and leg press max better than a 10- or 20-rep max, and the authors advise no more than 10 reps in this kind of equation.⁶
• In 103 young women in a required wellness course, formulas predicted the bench press max better from sets of fewer than 10 reps, before and after a training program.⁷
• In 69 college football players doing reps with 225 lb, 11 formulas were more valid when fewer reps were done.⁸
• In 118 male high school athletes, only 3 of 11 formulas gave bench press estimates not significantly off the tested max. The best, from a set of 7 to 10 reps, landed within 4.5 kg (about 10 lb) for 65% of them; about a third were further off.⁹
• In 20 trained lifters, half of them women, Epley's formula from a set to failure at 70% of the max came closer to the tested bench press max than two other formulas, with the least scatter of the four methods tested.¹⁰

People differ, too, and a formula gives everyone the same answer for the same set. A 2024 meta-regression of 952 tests to failure found that at 80% of a max, the spread between individuals (one standard deviation) was about 2.5 reps.¹¹ Treat the estimate as a starting point for your first heavy sets to correct.

How to use the number

• To set percentages. If a program says three sets of five at 80%, read the 80% row above and round to the plates or dumbbells you have; the warm-up set calculator builds the sets before it.
• To set loads by effort. The RPE table gives a first guess for a set at a target effort. Then let the set decide, as in how much weight to lift: adjust the next set by how many reps were left.
• To pick the reps and the effort. Which rep ranges fit muscle or strength is on how many reps; how hard each set should be is reps in reserve.
• To check progress. Re-estimate from the same lift and the same kind of set every few weeks; how to keep the number rising is progressive overload.

A review of the reps-in-reserve scale says its own percentage chart "should not be viewed as an absolute conversion tool", because people and days differ.¹ The same goes for every number on this page.

What it is not

It isn't a strength standard. The calculator estimates your own max from your own set. It doesn't say what anyone should lift for their body weight, age or sex, and it doesn't rank anyone. It also doesn't replace a tested max where a meet requires one.

Questions

Is the estimate as good for the squat and deadlift as for the bench press?

There's less evidence for them. The formulas use one curve for every lift, but lifts differ in how many reps people get at the same percentage. In 16 men, more reps were done at 60% of a max on the back squat than on the bench press or the arm curl,¹² and the 2024 meta-regression put 80% of a max at about 13 reps on the leg press and about 9 on the bench press.¹¹ Five of the six accuracy studies above tested only the bench press; the sixth added the leg press.^5,6,7,8,9,10 On a squat or a deadlift, estimate from a heavier set of 3 to 6 reps and let the first working sets correct it.

How does the calculator use RPE or reps in reserve?

The formulas assume the set went to failure. If it didn't, the calculator adds the reps you had left: 5 reps with 2 left counts as 7 to failure. On the lifting RPE scale, RPE 10 means no reps left, RPE 9 one, RPE 8 two, and so on,¹³ so RPE 8 adds 2 reps. Guesses run low: across 12 studies and 414 participants, a 2022 review found people underpredicted the reps they had left by about one, with better guesses in sets of 12 or fewer.¹⁴ So a guessed RPE tends to pull the estimate down, not up. Not sure? Leave it at none.

Epley or Brzycki: which is better?

Either, from a heavy set. Brzycki's did best from 10 reps or fewer in the largest study above, Epley's did best at 70% of a max in the smallest, and averaging two may reduce the error.^5,10 Hence the average.

Why does it stop at 12 reps?

Past 10, the estimates get loose: two studies above found the formulas less accurate there, and a third advises no more than 10.^5,6,7 The box takes up to 12 and warns above 10 reps to failure.