1-Rep Max Testing vs 1RM Formulas: Which Is More Accurate and What Does Each Cost You?

For most lifters, 1RM testing vs calculation is not a fair fight: a hard set of 3-5 reps run through a formula gets you a usable number at a fraction of the effort and risk. A true max test is more direct and, when repeated, quite reliable. It just rarely needs to be your default.

This article sits under our maximal strength and powerlifting science guide and covers the testing side: how accurate the formulas really are, what a max attempt costs, and how to do one sensibly when you need it.

One thing up front. You will often see claims that estimates land “within 1-3%” of a true max, and that a max attempt wipes you out for 7-10 days. We looked for the studies behind both and could not verify either number. The honest picture is messier, and more useful.

Is a true 1RM test actually reliable?

Yes, and this surprises people who assume a max is a noisy, mood-dependent number.

A 2020 systematic review pooled 32 studies (1,595 participants) on test-retest reliability. Intra-class correlations had a median of 0.97, and the median coefficient of variation (CV) was 4.2%, with individual studies ranging from 0.5% to 12.1%. The authors concluded the test has good to excellent reliability regardless of training experience, number of familiarization sessions, exercise, body region, sex or age (Grgic et al., 2020).

So if you test twice under similar conditions, you will usually land close to the same number. A median CV of 4.2% on a 150 kg squat is roughly 6 kg of day-to-day wobble. That is the “noise floor” every other method gets compared against.

How do 1RM formulas work?

Every rep-based formula rests on one observation: as the reps you can do at a weight go up, the weight you can lift for one rep goes down in a fairly predictable way. Do a hard set, count the reps, and work backwards.

The two you will meet most often:

Formula	Equation	Notes
Epley	weight x (1 + reps / 30)	Simple, tends to give higher values at low reps
Brzycki	weight x 36 / (37 - reps)	Becomes unstable above about 10 reps
Wathan	100 x weight / (48.8 + 53.8 x e^(-0.075 x reps))	Curved fit, lands close to Epley at 3-8 reps

Here is the same set run through both of the first two, as a worked example with a 100 kg bar:

Notice what this says. At 3 reps, Epley gives 110 kg and Brzycki 105.9 kg. Same set, same lifter. Neither is “wrong”. They are different curves fit to different data, which is why comparing your number to someone else’s calculator output is pointless.

You can run your own numbers in our free 1RM calculator.

How accurate are Epley and Brzycki?

Less exact than the internet suggests, but good enough for programming.

In a study of 220 men across several groups (untrained students, resistance-trained students, wrestlers, soccer and football players, and middle-aged trained men), six relative-endurance equations correlated with measured bench 1RM at r = 0.82 to 0.98. For sets of 10 reps or fewer, three equations significantly overpredicted and two significantly underpredicted, and the Brzycki equation was the most accurate. The authors cautioned that averaging two equations might reduce error (Mayhew et al., 1995).

A study in 70 adults aged 18-69 compared 1, 5, 10 and 20RM on the chest press and leg press. Prediction accuracy was best from the 5RM (R-squared of 0.993 on the chest press and 0.974 on the leg press), and the authors advised using no more than 10 reps in linear equations (Reynolds et al., 2006).

So where does “1-3%” come from? Probably from correlation values being misread as percentage error. A high correlation says the ranking of lifters is preserved. It does not say each person’s estimate is within 3% of their max. The same Reynolds study reported a standard error of about 3 kg on the chest press and about 16 kg on the leg press, in absolute terms. That is a useful reminder that error scales with the lift and with the person.

Do formulas work the same for women, trained lifters and different exercises?

Broadly yes, with caveats.

In 103 young women, bench press 1RM and repetitions to fatigue were measured before and after 12 weeks of training. 1RM rose 28% on average, but prediction equations were more accurate in both conditions when fewer than 10 reps were used. The authors concluded that training may alter the strength-endurance relationship without compromising the accuracy of predicting maximal strength (Mayhew et al., 2008).

For exercise choice, a 2024 meta-regression pooled 952 repetitions-to-failure tests from 7,289 people. It found little influence of sex, age or training status on how many reps are possible at a given percentage of 1RM, but more reps were possible on the leg press than on the bench press across the loading range (Nuzzo et al., 2024). The authors noted that more data are needed for other exercises, and that includes the squat and deadlift.

That last point matters for powerlifters. Most formula validation is bench and leg press. On squat and deadlift, the estimate is usually fine, but hold it a bit more loosely.

In the gym, the most common mistake is feeding a formula a “set of 5” where the last two reps were grinders with the hips shooting up or the bar bouncing off the chest. Count only reps you would show a judge.

What about velocity-based 1RM estimates?

Bar-speed devices are popular, so it is fair to ask whether they beat a simple rep formula. The evidence says not clearly.

A 2023 systematic review and meta-analysis of load-velocity 1RM prediction (26 studies, 20 with data in the meta-analysis, 434 participants) found moderate predictive validity, with a pooled standard error of the estimate of 9.8% (95% CI 7.4% to 12.2%). Predictions tended to overestimate, by about 4.5 kg or 3.7% of measured 1RM. The authors recommended direct 1RM assessment wherever possible, while allowing that velocity estimates may help with trend monitoring (Greig et al., 2023).

A 2024 study in 16 recreationally trained people reached a similar conclusion for the free-weight bench press. Most load-velocity regression equations were less accurate than an actual 1RM test, and including more loads, from light (around 20% 1RM) to heavy (80-90%), improved accuracy (Cabarkapa et al., 2024).

If you do not own a sensor, you are not missing anything essential.

What is the fatigue cost of 1RM testing?

This is the part of the question we can answer least firmly, so let’s be straight about it.

A near-max single is neurologically and mentally demanding, and anyone who has missed a heavy attempt knows a max day can feel flat afterwards. But we searched for controlled studies that tracked how long performance stays depressed after a true 1RM test in trained lifters, and did not find one we could verify. The popular claim of “7-10 days of systemic nervous fatigue” is not supported by a source we could check, so we are not repeating it as fact.

What we can say with more confidence:

• A max attempt is a high-effort, low-rep set. Total volume is small, so it is a very different stress from a long, hard training session.
• Powerlifters rated their 1RM attempts at about 9.6-9.7 on the RPE scale across squat, bench and deadlift (Helms et al., 2017). That is near-limit effort by definition.
• More attempts means more accumulated fatigue and more exposure. Keeping the attempt count low is the one lever fully in your control.

Individual response will vary with training age, the lift, sleep, food and stress. If you have just tested, keep your next session or two lighter and let performance, not a calendar, decide when to push again.

When is a true 1RM test worth it?

A true test makes sense when you need the real number, or when the lift itself is the goal.

• A meet is coming. You need to pick openers and know what you can do.
• A benchmark you care about. A once-or-twice-a-year check on a lift you want to track.
• A coach or programme asks for it. Some percentage-based programmes assume an accurate max.
• You are returning after a long break. Be cautious here: your old number is probably stale, so start well below it.

For everything else (a hypertrophy block, general strength, working out how heavy to go on Tuesday), a 3-5 rep set or an RIR-based top set does the job. Our guides on the RIR and RPE scale and on RPE-based top singles and back-off sets show how to train off effort instead of a fixed max. And if you wonder how well people judge effort, see how accurately lifters gauge RIR.

How to test a 1RM more safely

We did not find injury-rate data specific to 1RM tests in the studies we checked, so this is practical coaching advice rather than a research-proven protocol. The aim is to get an accurate number with as few heavy attempts as possible.

1. Pick the lift and keep conditions steady. Same bar, same time of day, same footwear and equipment as your previous test.
2. Warm up in a climbing sequence. Start with the empty bar, then climb in steps with fewer reps as the weight rises: for example 5 reps, 3 reps, 2 reps, then singles at roughly 70%, 85% and 92% of your best guess.
3. Rest properly between heavy attempts. Several minutes, not 90 seconds.
4. Plan the jumps before you start. Smaller jumps as you near the limit. Guessing in the moment is how people take a 10 kg jump they should not.
5. Cap it at about three max attempts. A fourth and fifth attempt are where technique tends to degrade.
6. Set up safety. A spotter, or safety arms in the rack for squat and bench. Learn how to bail before you need to.
7. Stop when form breaks. A grinder with a rounded back is not a PR, it is a risk.

Mental readiness matters more than people admit. Our piece on psychological readiness and arousal for max lifts covers the head side, and if you are building toward a meet, see the peaking and tapering guide.

Health note: if you have a heart or blood-pressure condition, are recovering from injury, or feel chest pain, dizziness or sharp joint pain when lifting heavy, see a doctor before testing maxes. Straining under maximal loads raises blood pressure, and this is a case where a number on a calculator is the better choice.

A realistic everyday workflow

Here is a simple way to combine both methods without overthinking it.

Week to week: do your normal top set at a load that leaves 1-2 reps in the tank. Log the reps. Once a month, run that set through your formula of choice and note the trend.

Every 6-12 weeks: do a 3-5 rep hard set on the main lifts and re-estimate.

Once or twice a year, or for a meet: test a true max.

If you train in an Indian commercial gym with crowded racks and no safety arms, this matters even more. When nobody is around to spot, the 3-5 rep set is the better choice almost every time. And if you want a ready-made structure to run these sets in, our free push, pull and leg day plans are a starting point.