How to estimate your 1RM from working sets, no max attempt

Your programme says five sets of three at 85 per cent, and you have no idea what to take 85 per cent of, because you have never tested a max. You do not need to. One working set, stopped at the right point, gives you a number accurate enough to plan around, without the risk or the three days of recovery a genuine max attempt costs.

What the number is actually for

A one-rep max is the heaviest load you can lift once with technique that still looks like yours, rather than a desperate scrape off the floor. You almost never use that figure directly. You use percentages of it, and percentages are all the planning really needs.

An estimate is usually written as e1RM. The distinction matters more than it sounds. A true 1RM is one measurement on one day, with a belt, a spotter and a shot of adrenaline behind it. An e1RM comes out of sets you were doing anyway, most weeks, so you end up with a trend rather than a single data point. If you do want to test properly at some stage, we covered how to do that safely in its own piece.

The method: one set stopped short of failure

Warm up as you normally would, pick a load you expect to get for three to eight reps, and take the set to the point where one or two reps are still in the tank. That reserve is the whole point. You are not going to failure; you stop while you are confident another clean rep is there.

Write down three things: the load, the reps you completed, and how many you left. If you reckon one was left, put one extra rep into the formula. If two, add two. Skip that step and your estimate drifts low every single time.

The three-to-eight window is not arbitrary. Below three reps the load is already so close to a max that you have lost the reason for avoiding a test in the first place. Above eight or ten, conditioning and how much discomfort you tolerate start deciding the rep count, and the estimate turns to mush. Three to six is the sweet spot.

Two formulas and a percentage chart

The best known is Epley: e1RM = load × (1 + reps ÷ 30). For 100 kg for five reps, that is 100 × 1.167, so roughly 117 kg.

The other is Brzycki: e1RM = load ÷ (1.0278 − 0.0278 × reps). The same input gives about 112 kg. Five kilos of disagreement on identical data is the clearest possible reminder that this is an estimate, not a measurement.

If you would rather not do arithmetic between sets, a percentage chart does the same job:

• 1 rep — 100 per cent
• 2 reps — about 95 per cent
• 3 reps — about 92 per cent
• 4 reps — about 89 per cent
• 5 reps — about 86 per cent
• 6 reps — about 83 per cent
• 8 reps — about 78 per cent
• 10 reps — about 73 per cent

Read it backwards. Five reps with 100 kg means that 100 kg was roughly 86 per cent of your max, so the e1RM is around 116 kg. Chart and formula will not agree to the kilo, and that is fine.

Where the estimate falls down

These formulas were built around big barbell lifts and that is where they behave. Squats and deadlifts estimate more reliably than the bench press, where a small change in technique shifts the numbers a lot. On isolation work such as curls or lateral raises, estimating a max serves no purpose and you can leave it alone.

Then there is the individual spread. Two lifters with the same genuine max might get six and twelve reps respectively at 80 per cent. If you are the kind of lifter with a lot of reps in you, the formula will overstate your max; if you are explosive but fade quickly, it will understate it. The number is yours, for that lift, and comparing it with someone else's tells you very little.

Judging reps in reserve is also a skill. Newer lifters routinely believe one rep is left when three or four actually are. That drags the e1RM down, which is not dangerous but does mean months of training lighter than you could. The odd set taken to genuine failure on a machine calibrates the feel quickly.

Finally, there is the day itself. Sleep, food, stress and the workload of the past fortnight will move an e1RM by five per cent in either direction with no change in your actual strength.

Using the number week to week

The value here is not what the figure says today, it is that you have one constantly. Take one working set per lift, usually the last and heaviest, and run the numbers. Over eight weeks that gives you a line, and a line is far more honest than any single session.

Programme your percentages from the e1RM, but if the sets keep landing far too light or far too heavy, adjust the load based on what happens under the bar rather than what the chart says. The formula is the starting point; the set is the final word. The choice between heavier loads and higher reps also changes how these numbers move, which we go through in a separate piece.

If the e1RM sits flat for six weeks or more, that is a programming signal rather than a cue to try harder — the ways out of a strength plateau are worth working through. And if you are coming back from time off, recalculate from scratch instead of trusting your old figure; the first six weeks back follow their own logic.

When to see a doctor

Estimating is safer than testing, but it still means straining under a heavy load. See a doctor or physiotherapist for pain lasting more than two or three weeks, pain that shows up at rest or at night, or numbness and weakness in an arm or leg. Chest pain, dizziness or breathlessness out of proportion to the effort means stopping the session there and then. If you have a known heart condition, high blood pressure or a hernia, have that conversation before your first heavy set rather than after it.

The short version

• Take a set of three to eight reps and stop with one or two in reserve; record the load, the reps and the reserve.
• Epley: load × (1 + reps ÷ 30). Brzycki: load ÷ (1.0278 − 0.0278 × reps). Expect them to disagree slightly.
• Estimates are sharpest at three to six reps on big barbell lifts; do not bother on isolation work.
• Allow for roughly five per cent of error either way from the lift, the day and your judgement of reps in reserve.
• Follow the trend across weeks rather than any single figure, and treat it as a starting point for percentages, not a rule.