Calculator
Your one-rep max (1RM) is the most you can lift once with good form, and you don't need to test it to know it. Enter a weight you lifted for 3 to 10 reps and how many more you could have done, and this calculator estimates your 1RM as the average of two classic formulas, Epley and Brzycki, with five more alongside. Use the result to pick starting weights, then let reps in reserve (RIR) fine-tune them.
Enter the weight and reps from one hard set, pick how many more good reps you had left, and choose the lift if you want its own reps table and cautions.
75kg
For most people the estimate is close, but it can be 10% or more out: in studies using sets of 10 reps or fewer, individual errors had a standard deviation of about 9 to 11%.
The seven formulas range from 72 to 76.5 kg, with w the weight and r the reps to failure (8).
| Formula | Estimate |
|---|---|
| Epleyw × (1 + 0.0333 × r) | 76 kg |
| Brzyckiw ÷ (1.0278 - 0.0278 × r) | 74.5 kg |
| Landerw ÷ (1.013 - 0.0267123 × r) | 75 kg |
| Lombardiw × r^0.10 | 74 kg |
| Mayheww ÷ (0.522 + 0.419 × e^(-0.055 × r)) | 76 kg |
| O'Connerw × (1 + 0.025 × r) | 72 kg |
| Wathenw ÷ (0.488 + 0.538 × e^(-0.075 × r)) | 76.5 kg |
| % of max | Weight | Reps to failure |
|---|---|---|
| 95% | 72.5 kg | about 3most people 2 to 5 |
| 90% | 67.5 kg | about 5most people 3 to 7 |
| 85% | 65 kg | about 7most people 5 to 9 |
| 80% | 60 kg | about 10most people 7 to 12 |
| 75% | 57.5 kg | about 12most people 10 to 15 |
| 70% | 52.5 kg | about 15most people 12 to 18 |
| 65% | 50 kg | about 17most people 13 to 21 |
| 60% | 45 kg | about 20most people 15 to 24 |
Use these as starting weights. On big lifts, stop 2 to 3 reps short of failure; on the last set of small moves you can go to 0 to 2. In your first two weeks, leave 3 or 4. Most people misjudge reps left by about one. Thinking of testing a true max? Check who shouldn't.
Lift a weight you can manage for 3 to 10 reps, enter the reps and how many you had left, and the calculator estimates your max. Estimates were most accurate with 10 reps or fewer, and heavier sets of about 3 to 8 reps did best where rep ranges were compared.
The most accurate formula changed with the lift, sex and group tested. Between 3 and 10 reps the seven classic formulas give answers within about 6 to 8% of each other (our arithmetic), less than the error for one person, so the headline is the average of Epley and Brzycki, with all seven underneath. One study suggests averaging two formulas may reduce the error.
Most formulas were right on average to within a few per cent, but individual errors had a standard deviation of about 9 to 11%. Every formula underestimated older adults on machines and untrained students' deadlifts, and estimates from the leg press can run high, because people manage more reps there at the same percentage of their max.
Across 269 studies, people managed about 10 reps at 80% of their max (about 9 on the bench press and 13 on the leg press), with a spread of 2 to 4 reps between people. The calculator's table of starting weights uses these figures, not an old coaching chart.
Reps to failure = reps done + reps left, and each choice of reps left shows its rating of perceived exertion (RPE): 10 means none left, 9 means one. Most people underestimate the reps they have left by about one, so estimates from sets stopped short tend to come out slightly low (our reading), which is the safe direction.
Use the calculator instead if you're new to lifting, pregnant, recovering from heart or hernia surgery, or have uncontrolled blood pressure, heart disease without your care team's agreement, or diabetic eye disease. In testing studies, injuries were most common in older people new to lifting.
Every formula works from the weight you lifted (w) and your reps to failure (r): reps to failure = reps done + reps left. So 8 reps with 2 left in the tank counts as 10.
Reps done plus reps left can't come to more than 12, and at 11 or 12 the calculator says the estimate is less accurate and suggests a heavier weight. Above 10 reps the formulas drift apart, and the studies found estimates most accurate with 10 or fewer. With 1 rep to failure, the weight you lifted is your max.
Written as the two full-text validation studies printed them, with Mayhew's matching its original abstract (e is Euler's number, about 2.718). The headline is the average of the first two; the others are under "See every formula", with their range. Some published copies carry typos, such as Epley with 0.333 or Brzycki with a plus sign, so check any formula you copy against these.
Most weren't fitted to published data. Epley's began as "a poundage chart, not based on scientific research", Lander's as a "guess-timated" chart, and Brzycki's paper doesn't say what it was based on. Mayhew's is the exception: it was fitted to one-minute rep tests at 55 to 95% of the max in 184 men and 251 women.
Epley and Brzycki give the same answer at 10 reps; below 10, Brzycki gives the lower one. From 3 to 10 reps the seven agree within about 6 to 8%, and the average of Epley and Brzycki stays within about 1 to 2% of the average of all seven. The gap widens to about 13% at 12 reps and 25% at 15.
At 1 rep, Epley would give 103% of the weight and Mayhew 109%, because Mayhew's curve was fitted to one-minute rep tests rather than single lifts. That's why the calculator gives the weight itself for 1 rep to failure.
An example: 60 kg for 8 reps to failure gives Epley 76.0 kg and Brzycki 74.5 kg, so the estimate is 75 kg (their average, 75.2, to the nearest half kilo). The other five give 72.0 to 76.6 kg. The same set stopped with 2 reps left counts as 10 reps to failure and gives 80 kg.
The table under the result takes 95% down to 60% of your estimate in 5% steps, rounded to 2.5 kg or 5 lb, with the reps most people manage at each from the 269-study table: the bench press column for the bench press, the leg press column for the leg press, and the main table for every other lift.
In 70 adults on the chest press and leg press, a 5-rep max predicted the one-rep max best, explaining 99% and 97% of the variation, against 96% and 92% from a 20-rep max; the authors advise that "no more than 10 repetitions should be used". In 103 women the formulas were more accurate with fewer than 10 reps to failure, and over the full range of 2 to 30 reps Brzycki's and Lander's formulas were out by 27% and 23% on average, with standard deviations of 102% and 71%. Validity was also higher with fewer reps in 69 college football players, and trained older adults predicted best from sets of 5 to 8.
In 119 young adults, Lombardi's formula was the best for men's bench press and squat, Brzycki's and Lander's were among those that matched women's bench press, and none was accurate for men's arm curl. In 67 untrained students, only 2 of 7 formulas matched the tested bench press on average, and only 1 the squat. In 220 men of different training levels, Brzycki's was the most accurate of six with 10 reps or fewer, and the authors suggest that averaging two formulas may reduce the error. In 20 resistance-trained adults lifting 70% of their max, Epley's had the smallest error of three, and in 18 people with knee osteoarthritis the best formulas differed between the knee extension and the leg press.
Between 3 and 10 reps the seven formulas here differ by about 6 to 8% (our arithmetic), which is less than the error for one person. That's why the headline averages two and the rest are shown underneath.
In the 46 women who did 10 reps or fewer, the seven formulas here were out by 4% or less on average, but individual errors had standard deviations of about 9 to 11%: on bench presses averaging about 28 kg, roughly 3 kg either way for about two in three women (our arithmetic). In 119 young adults, the limits of agreement for the better formulas ran from about 2 to 4 kg either way on bench presses of 29 to 69 kg to about 10 kg either way on squats of about 117 kg. At worst, Brzycki's and Lander's formulas overestimated men's squat by about 14 kg on average.
In 67 untrained students, every one of seven formulas underestimated the deadlift. In 51 adults averaging about 71 on 11 gym machines, every formula underestimated the tested max, with correlations of 0.77 to 0.90 for upper-body lifts and 0.60 to 0.80 for lower-body lifts.
People manage more reps on the leg press at the same percentage of their max: about 13 at 80%, against about 9 on the bench press. A formula reads those extra reps as a heavier max, so for someone who does the average leg press reps at 80%, Epley and Brzycki come out about 15 to 20% high; on the main table the same sum gives about 6%, and on the bench press about 2 to 3% (our arithmetic). Endurance runners also managed far more leg press reps at 70% of their max than weightlifters (39.9 against 17.9), so formulas may overestimate the max of people with an endurance background (our reading).
In 103 women tested before and after 12 weeks of training, the formulas were more accurate with fewer than 10 reps both times. With 2 to 10 reps, 9 of the 14 formulas tested matched the tested max before training and 5 after, and the authors concluded that training didn't compromise the predictions. In 119 young adults tested before and after 16 weeks, the authors note that some formulas worked as well for trained people as for beginners.
Across 269 studies of 7,289 people, people managed 9.8 reps on average at 80% of their max (a between-person standard deviation of 2.5), 4.9 at 90% and 14.8 at 70%. On the bench press it was 8.8 at 80%, and on the leg press 13.1. Sex, age and training status didn't clearly change the relationship. The traditional coaching chart is close at heavy loads and for the bench press, but lists fewer reps than people manage at lighter loads, most of all on the leg press.
In 12 studies with 414 people, people tended to underpredict their reps to failure, by 0.95 reps on average, and they judged better close to failure and in sets of 12 reps or fewer. In 141 trainees the prediction errors had a standard error of measurement of 2.6 to 3.4 reps, with a tendency to improve with experience. On the RIR-based effort scale, RPE 10 means no reps in reserve and RPE 9 means one; it was tested in experienced and novice squatters.
The American College of Sports Medicine (ACSM) says training to failure isn't necessary, that it may be inadvisable for older people, and that stopping 2 to 3 reps short gives enough effort. A properly run 1RM test is reliable (in 32 studies, repeat tests typically varied by about 4%), but loads take 2 to 3 practice sessions to settle, and the authors of one validation study say testing may not be prudent with joint discomfort or injuries.
The calculator's table of starting weights uses the research table, not the coaching chart.
Average reps to failure at each percentage of the one-rep max, from Nuzzo and colleagues' 2024 analysis of 269 studies. The first figure covers every lift except the bench press and the leg press, which have their own. In brackets is the range for most people: the average plus or minus one standard deviation between people, which covers roughly two in three (our arithmetic, rounded to whole reps).
At 80% of your max, most people manage about 10 reps, about 9 on the bench press and about 13 on the leg press. The spread between people grows as the weight gets lighter and the estimates get less certain, so the table stops at 60% (our choice). Most of the data came from healthy people aged 20 to 40. The squat, looked at separately, sat close to the main table: about 9.7 reps at 80% and 5.0 at 90%.
Choose the bench press and the table uses the bench press figures; choose the leg press and it uses the leg press figures. Every other choice, including other upper-body lifts, squats, deadlifts and machines, uses the main table, which the authors say applies to all exercises other than the bench press and leg press.
Source 9
This chart was printed in a commonly assigned strength and conditioning textbook for years. It was presented as a general guideline based on a small number of studies, gives no spread between people, and its references are largely not peer reviewed. Compared with the research table it is close at heavy loads and for the bench press, but lists fewer reps than people manage at lighter loads, most of all on the leg press.
In 16 men, more reps were done in the back squat than in the bench press or arm curl at 60% of their max, with smaller differences at 80% and 90%. Endurance runners did 39.9 leg press reps at 70% of their max against 17.9 for weightlifters, and 19.8 against 11.8 at 80%, but no more at 90%. In a 1990 study, trained and untrained men differed for the arm curl, knee extension and sit-ups, but not the bench press or leg press.
Treat the weights as starting points. If you want sets of 8 with 2 reps to spare, pick the row where most people manage about 10, do a set, and adjust by how many reps you had left. In the ACSM's review, muscle grew about as much with loads from 30% to 100% of the max when sets were hard, while heavier loads, 80% or more, built more strength.
Reps in reserve (RIR) is how many more good reps you could have done when you stopped a set. On the RIR-based effort scale, an RPE of 10 means no reps left, 9 means one, 8 means two, and so on, which is how the calculator labels its "reps left" choices.
Source 11
The calculator's effort note matches our training guides:
Most people underestimate the reps they have left by about one, so a set reported as "2 left" probably had about 3. Each missed rep lowers the estimate by about 3% at typical rep counts (our arithmetic), so estimates from sets stopped short tend to come out a little low. For choosing training weights, that's the safe direction.
For the most accurate estimate, use a set of about 3 to 8 reps that ended within 1 or 2 reps of failure, since people judge reps left better close to failure. Beginners and older lifters can use a set stopped 2 to 3 short instead. Don't grind to true failure on squats or bench presses without a spotter or safety arms; the authors of one validation study advise spotting on both.
Supervised one-rep max tests have a good record in many groups. The American Heart Association reports no significant heart events during resistance training and 1RM testing in healthy adults, low-risk heart patients, people with controlled high blood pressure, people after a stroke and transplant recipients. Across 23 trials in coronary heart disease (1,174 people), there was 1 heart complication during resistance training and none during strength testing, against 63 during aerobic training and testing.
Injuries are more common in older and inexperienced lifters: testing injured 2 of 83 adults aged about 66, both without lifting experience, and caused 11 injuries in 57 people aged 70 to 79. Heavy lifts to failure with a held breath also drive blood pressure very high: 320/250 mmHg on average on a double leg press in 5 bodybuilders, and above 480/350 in one.
Our list, from the guidance in this page's sources:
This is how the validation studies tested, not a plan to follow: a warm-up, then 3 to 5 attempts with at least 3 to 5 minutes' rest between them, adding about 3 to 10% each time, through a full range of motion, with a spotter on the bench press and squat. Loads took 2 to 3 practice sessions to settle, more for complex lifts and beginners. Breathe out on the effort rather than holding your breath. Done this way the test is reliable: in 32 studies, repeat tests typically varied by about 4%.
The barbell workout uses sets of five on the squat and bench press. The technique sessions build each lift up from bodyweight and light weights before any heavy set.
Barbell strength workout A: back squats and bench presses in sets of five, plus rows. Workout B (squat, overhead press, deadlift) is a separate session.
40 minBeginner6 movesA band, A barbellA 25-minute practice session for the squat: find your depth on a chair, groove the bodyweight squat with pauses, then load it with a slow goblet squat.
25 minBeginner8 movesA wall, A bench or boxA 30-minute practice session for the bench press: shoulder blade work, push-ups and dumbbell presses, then the empty bar with the safety arms set.
30 minBeginner7 movesA band, A dumbbellA 30-minute session from the hip hinge to your first deadlift: hinge practice, kettlebell deadlifts, slow Romanian deadlifts, a light bar and a carry.
30 minBeginner9 movesA kettlebell, A dumbbellPick a weight you can lift 3 to 10 times with good form, do one set, and note the reps and how many more you could have done. Enter them in the calculator: it adds the reps you had left and averages the Epley and Brzycki formulas. Estimates are most accurate with 10 reps or fewer.
None wins every time. Brzycki's was the most accurate with 10 reps or fewer in one study of 220 men, but in others the best formula changed with the lift, with sex and with the group tested. Below 10 reps the seven common formulas give answers within about 6 to 8% of each other, so we show the average of Epley and Brzycki, with all seven under "See every formula".
About 3 to 8 is best, and 10 or fewer is fine. In 70 adults a 5-rep max predicted the one-rep max best and the researchers advised no more than 10 reps, and trained older adults did best with sets of 5 to 8. Above 10 the formulas drift apart quickly, so the calculator flags 11 or 12 reps to failure and won't go higher.
Yes, with caveats. Every formula underestimated untrained students' deadlifts in one study, so your deadlift max may be higher than the estimate. People manage more reps on the leg press at the same percentage of their max, which pushes estimates up, so use a heavier set of 5 reps or fewer there. Choose the lift in the calculator to see its caution and, for the leg press, its own reps table.
Yes. Choose how many more good reps you could have done in the "reps left" box; each choice shows its RPE (rating of perceived exertion), where 10 means no reps left, 9 means one, and so on. Most people underestimate reps left by about one, so the estimate may come out a little low, which is the safe direction.
About 10 on average across 269 studies, about 9 on the bench press and about 13 on the leg press, with most people within 2 to 4 reps of that. At 90% it's about 5 (about 4 on the bench press), and at 70% about 15. Other upper-body lifts follow the main table, which covers every lift except the bench press and leg press.
For strength, heavier sets of about 5 to 8 reps, roughly 80 to 90% of your max on the bench press; for muscle, 8 to 12 reps works for most moves. Muscle grows across a wide range of loads when sets are hard, so use the percentages to choose a starting weight and let reps in reserve fine-tune it. Our sets and reps guide has the detail.
You don't need to. Supervised tests have a good safety record in healthy adults, but injuries were more common in older people new to lifting, and heavy lifts to failure drive blood pressure very high. Skip a true max if you're new to lifting, pregnant, recovering from heart or hernia surgery, or have uncontrolled blood pressure, heart disease or diabetic eye disease.
It works, but expect a little more error. In adults around 70 on gym machines every formula underestimated the tested max, so tick "60 or over" to see that caution, and trained older adults predicted best from sets of 5 to 8. Beginners get stronger quickly, so re-check every few weeks.
Strength varies from day to day, and so does how close to failure you take a set. Fatigue, sleep, your warm-up and how many sets you've already done all change the reps you manage. Compare estimates from the same lift at the same point in your session.
Estimates for healthy adults, not a personal prescription or a diagnosis, and clinical review of these pages is pending. Check with a GP before hard training or testing a maximum if you have a heart condition, high blood pressure, chest pain, dizziness or breathlessness when active, and stop for sharp or increasing pain. Health disclaimer.
