Estimating your one-rep max without testing it

Testing a true one-rep max is disruptive. It requires a peaking period, a spotter, full recovery afterwards, and a willingness to grind a maximal attempt with imperfect technique. For most people training for size or general strength, it costs more than it returns.

So we estimate instead, from a set taken close to failure. The question is how much confidence that estimate deserves.

What the estimate is for

A one-rep max is useful mainly as a programming anchor. Percentage-based programmes need a reference number, progress tracking needs a common unit across rep ranges, and comparing a five-rep set today against an eight-rep set last month requires converting both to something.

None of those uses require the number to be exactly right. They require it to be consistently wrong in the same direction — which is a much easier standard to meet.

Five formulas, five answers

The common equations were each fitted to a different sample, with different lifts and different training populations. That is why they disagree.

• Epley — the most widely used; tends to read high at low rep counts.
• Brzycki — reads lower as reps climb; a linear inverse relationship.
• Lombardi — a power-curve fit, behaving differently at the extremes.
• Wathan — an exponential fit that tends to track well across the widest rep range.
• O’Conner — a simple linear approximation, useful mainly for mental arithmetic.

Feed the same set into all five and you typically get a spread of a few percent. That spread is not a defect to be resolved by picking a favourite. It is an honest depiction of how much uncertainty exists in predicting a maximal effort from a submaximal one.

Where the estimates break down

Rep count is the biggest factor. Estimates from two to six reps are reasonably trustworthy. Beyond about ten, accuracy falls away quickly, and beyond twelve the formulas diverge enough that the answer is barely worth calculating. The reason is that high-rep sets are increasingly limited by fatigue, breathing and technique breakdown rather than by maximal force production — the thing a 1RM actually measures.

Most of these formulas were never derived from research. This is the part worth sitting with. Brzycki’s appeared in a practitioner article that never states what data it was fitted to. ¹ Epley’s began life as a poundage chart in a training manual ², Wathen’s as a load-assignment table in a coaching textbook ⁴, Lombardi’s in an introductory weight training book ⁶. Cross-validation studies that later tested them note plainly that most gave no evidence of the population used to develop them or how they were derived statistically — the formulas were extrapolated from the charts afterwards.

That does not make them useless. They were built by experienced coaches from observed loading practice, they broadly agree with each other, and when tested against real maxima they correlate well. But it does mean the precision they project is borrowed. A number carried to one decimal place, from an equation reverse-engineered out of a 1985 wall chart, is not a measurement.

The lift matters too. Estimates tend to be more accurate for movements with a longer range of motion and a clear failure point. The most-cited comparison of these equations tested seven of them across the bench press, squat and deadlift, and found that every single one underestimated the deadlift. ⁵ The correlations were high across the board — above 0.95 — which is exactly the kind of number that hides a consistent absolute error. If you are estimating a deadlift max from reps, expect the figure to read low.

Our calculator does not ask which lift you are doing, so it cannot correct for this. Treat a deadlift estimate as a floor.

Training history matters — and cuts both ways here. The study that introduced the reps-in-reserve scale compared experienced and novice squatters directly and found experienced lifters better at judging how close to failure they were. ⁸

The picture is not settled, though, and it is worth saying so. A later meta-analysis pooling twelve studies and 414 participants found no overall effect of training experience on how accurately people predict reps to failure. It did find something more consistent: across the board, people under-predict — they typically have about one more repetition in them than they think. So if you stop at what feels like two reps in reserve, you were probably closer to three.

That same validation study used 67 untrained college students, so the equations most people rely on were checked against beginners rather than experienced lifters. ⁵ Novice lifters often cannot express a true maximum because they lack the neural efficiency and confidence to do so. Their estimates commonly read higher than what they could actually lift.

Proximity to failure matters most of all. Every one of these formulas assumes the set was taken to, or very near, failure. A set with four reps left in the tank produces an estimate that is simply wrong, and no formula can detect that from the numbers you enter.

A more practical alternative

For day-to-day programming, RPE and reps-in-reserve often beat percentages outright.

The logic is straightforward. A percentage of an estimated max is a fixed prescription applied to a body whose readiness changes daily — with sleep, stress, food and accumulated fatigue. Eighty percent on a good day and eighty percent after a bad night are the same number and very different sets.

Reps-in-reserve inverts this. Instead of “five reps at 80 percent”, the instruction is “five reps leaving two in the tank”, and the load self-adjusts to how you are actually performing. Research validating the RIR-based RPE scale found that experienced lifters estimate proximity to failure reasonably well, particularly within a few reps of it.

Our calculator does both: it gives you the percentage table if you want it, and converts reps plus RIR into a working load if you would rather train that way.

The bottom line

Estimate from a set of two to six reps taken close to failure, look at the range rather than fixating on one formula’s answer, and re-estimate every few weeks rather than trusting a number from three months ago.

Then, for the actual training, consider ignoring the percentage entirely and prescribing by how many reps you left in reserve. It is a more honest instruction, because it accounts for the one variable no formula can see: how you feel today.