Why the Percentage Chart Isn't Always Right
The chart says five reps is 87 % of your max. For some lifters it is 80 %, for others 92 %. That spread is not an error in the chart — it is what a chart is — and understanding it makes the chart more useful, not less.
A percentage chart is a useful tool. It is also an average, and an average describes a population, not you. This article is about where the chart fails and what to do about it.
Worth saying up front: this site maintains its own percentage chart. Explaining its limits is not a contradiction — a calculator is useful exactly to the extent that you know what it can and cannot do.
Where the numbers come from
The relationship between reps and percentages is not a measured constant of nature. It is an equation fitted to the performance of test subjects, and there are several competing equations. The best known — Epley, Brzycki, Lombardi, Wathan and O'Conner — were published in the 1980s and 1990s and rest on fairly small datasets.
Some of them, Epley and Brzycki among them, were originally published as coaching rules of thumb with no subject data at all. They have since been evaluated in studies, and it is those studies the following three features describe: the subjects were often students rather than competitive lifters, the test lift was typically a bench press or a leg press, and the sample sizes were in the dozens rather than the thousands.
So the equations describe how a particular kind of population behaved in particular lifts. Apply the same equation to an experienced powerlifter pulling a deadlift and you are outside the data.
The formulas disagree — and the disagreement grows
The clearest evidence of the chart's uncertainty is that the formulas do not agree with each other. The 1RM calculator runs all five side by side, so you can see the spread directly.
Take 100 kg and look at what the five formulas estimate as a max from different rep counts:
| Reps | Lowest estimate | Highest estimate | Spread |
|---|---|---|---|
| 3 | 105.9 kg | 111.6 kg | 5.7 kg |
| 5 | 112.5 kg | 117.5 kg | 5.0 kg |
| 8 | 120.0 kg | 127.7 kg | 7.7 kg |
| 12 | 128.2 kg | 144.0 kg | 15.8 kg |
| 15 | 131.1 kg | 163.6 kg | 32.5 kg |
At three reps the formulas essentially agree: a six-kilo band around 110 kg is accurate enough to program with. At fifteen reps the band is 32.5 kg. That is not an estimate any more, it is an opinion.
Which gives the first practical rule: the closer the set is to a single rep, the more trustworthy the estimate. A set of three or five is a good source. A set of twelve is not.
This site averages all five. That beats picking one formula arbitrarily, because the individual extremes cancel each other out. It does not make the underlying problem smaller — an average of uncertain estimates is still an uncertain estimate.
The lift matters more than the chart admits
Here is the chart's single biggest shortcoming: it gives the same answer regardless of which lift you are doing. In practice, rep tolerance varies between lifts systematically.
Deadlifts tolerate fewer reps. The load is large, the range of motion is long, and every rep starts from a dead stop. Lifters typically get fewer reps at a given percentage than the chart promises. Many get three reps with a load the chart says should give five.
Squats and bench press sit closer to the chart. They are also the lifts the formulas were derived from, so this is no surprise.
Machines tolerate more reps. When balance and stabilisers are not the limit, the set ends when the target muscle tires. A leg press at 80 % can produce considerably more reps than the chart suggests.
The practical consequence: do not apply one chart to every lift without checking. In the deadlift especially, treat the chart's rep counts as an upper bound.
Lifters are different
The second large source of spread is the lifter. Two people with the same 100 kg max can get different rep counts at 80 % — and the gap can be a factor of two.
Several things sit behind this. Muscle fibre composition affects how quickly force falls off across reps. Neural capacity to express maximal force varies, and it affects the single rep specifically while barely touching a set of ten. Body proportions and leverages decide where in the range of motion the lift breaks down.
In practice lifters fall roughly into two types. Some have a high single relative to their rep capacity — a strong one-rep max, but they fatigue fast. Others are the reverse: a more modest single, but an unusual number of reps at the same percentage.
Neither is better. What matters is that one chart serves those two lifters differently, and both are better off knowing which end they sit at.
Your chart moves with your training
This is the part that surprises most people. Rep tolerance is not a fixed trait but a trainable skill, and it follows whatever you have been doing lately.
Train for months at low reps and heavy loads, and your one-rep max rises more than your ability to grind out sets of ten. By the chart's measure you become "bad" at reps. Spend blocks in the 8–12 range and the opposite happens.
Which leads to something worth sitting with for a moment: your personal percentage chart shifts over a training block. The relationship you measured in week one may not hold in week twelve, even at the same max.
The max is an estimate too
One more source of uncertainty deserves naming. The percentage is calculated from a max, but that max is rarely a fresh measurement. It is either an estimate from a work set — the output of a formula — or a test done weeks ago.
If the max is itself five percent off, every percentage derived from it is five percent off. The errors do not cancel, they compound. This is why RPE works better for many people: it never needs a max as an input at all.
Building your own chart
The answer is not to abandon the chart but to calibrate it to yourself. It takes a few weeks of note-keeping and no extra training.
- Record the reps whenever a set goes close to the end. Any time you hit RPE 9 or 10, write down the lift, the load and the reps.
- Work out the actual percentage. Divide the load by your current max estimate. Now you have one data point: "deadlift, 5 reps, 84 %".
- Compare with the chart. The chart says 86.8 % for five reps. If your observations sit consistently below that, your rep tolerance in that lift is lower than the chart's.
- Correct per lift. After a few weeks a pattern appears. Write down your own correction — "deadlift: chart minus 3 percentage points" — and program with it.
Ten or so observations per lift turns out to be surprisingly sufficient. After that the chart is not an average any more, it is your chart.
When the chart is fine as it is
Not everything needs calibrating. The chart is accurate enough when:
- You are calculating warm-up sets. A few percent of error at 60 % means nothing.
- You are sketching the rough shape of a block and intend to adjust loads along the way anyway.
- The reps are 1–5 and the lift is a squat or a bench press.
- You are comparing your own numbers across time using the same method. A systematic error cancels out when both figures were produced the same way.
Accuracy matters when the load is close to maximal, when there are many sets, or when a program leans on a percentage for weeks without checking in.
Summary
A percentage chart is an average drawn from small datasets, collected from other people doing other lifts. It is a good starting point and a poor final authority.
Three things to carry: trust the chart most in the 1–5 rep range, assume the deadlift tolerates fewer reps than it promises, and log your own completed sets for a few weeks. After that you have a chart that describes you rather than a group of 1980s test subjects.
The starting point is in the percentage chart, and for adjusting to the day, the RPE chart.