How to Increase Your 1 Rep Max: The Arithmetic of Getting Stronger

Every method for increasing a one-rep max is a way of organising the same thing: asking the body to handle slightly more than it is used to, then recovering. That part is not controversial. What usually goes missing is the arithmetic, and the arithmetic changes what you should do. It tells you when an extra rep beats an extra plate, how much headroom a normal strength block leaves, how often a realistic lifter should expect to set a PR, and how far your programmed percentages drift if you stop retesting. All the figures below are computed, not quoted, and the assumptions are listed at the end.

A rep PR versus a plate PR

Progress does not have to mean more weight on the bar. Doing the same weight for one more rep is a PR too, and the two are not worth the same amount. Using the Epley formula, 1RM = w(1 + r/30), an extra rep at a fixed weight always adds exactly w/30 to the estimate, whatever the rep count. An extra 2.5kg at a fixed rep count adds 2.5(1 + r/30). Setting those equal gives the crossover weight:

w = 75 + 2.5r

So at 5 reps the crossover is 87.5kg. Below that, add weight. Above it, the rep is worth more than the plate. For a 100kg working weight the numbers look like this:

Reps at 100kg	Est. 1RM	+1 rep adds	+2.5kg adds	Better move
3	110.0kg	+3.33kg	+2.75kg	Rep
5	116.7kg	+3.33kg	+2.92kg	Rep
8	126.7kg	+3.33kg	+3.17kg	Rep (just)
10	133.3kg	+3.33kg	+3.33kg	Equal

At 100kg for 10 the two moves are worth exactly the same, because 75 + 2.5 × 10 = 100. A lifter benching 60kg is on the other side of the line at every rep count: for them a 2.5kg jump at 5 reps is worth 2.92kg of estimated max and an extra rep only 2.00kg. This is the real reason the advice differs for beginners and intermediates. Beginners should chase weight because their weights are small enough that a plate is the bigger step. Once the bar is heavy, reps at the same weight are the more efficient PR, and they are also the lower-risk one. The Brzycki formula, which our calculator also runs, values reps more steeply still: at 100kg it rates the sixth rep at +3.87kg and the ninth at +4.76kg. See Epley vs Brzycki for why the two disagree.

What 2.5kg actually means at your weight

The smallest jump in most UK gyms is 2.5kg (a 1.25kg plate each side). That jump is not the same size for everyone:

Current max	2.5kg as a %	What it is like
40kg	6.25%	A month of a beginner's progress in one jump
60kg	4.17%	Still a large step
100kg	2.50%	Roughly one training block
140kg	1.79%	A quarter of an intermediate year
180kg	1.39%	Half an advanced lifter's year
220kg	1.14%	Fractional plates become essential

This is why linear progression works for beginners and stops for everyone else. Adding 2.5kg every session, three times a week, is 90kg in twelve weeks. On a 60kg squat that is a 150% increase in three months, which is impossible, so nobody actually sustains it. Even 2.5kg a week is 130kg a year. The rate at which you can add weight is a percentage of your max, not a number of plates, and the percentage falls as you get stronger. Past the first year the answer is to stop progressing session to session and progress block to block instead.

A strength block, with the tonnage worked out

The standard four-week intensification block, written off a 100kg max, is the simplest version of block-to-block progression. It is worth seeing what it does numerically rather than just following it:

Week	Sets × reps	Intensity	Load	Total reps	Tonnage
1	5 × 5	75%	75kg	25	1,875kg
2	4 × 4	80%	80kg	16	1,280kg
3	3 × 3	85%	85kg	9	765kg
4 (deload)	3 × 5	60%	60kg	15	900kg
5	Heavy single or estimated retest, then rebuild the block from the new number

Intensity rises 10 percentage points across three weeks while total work falls by 59%, from 1,875kg to 765kg. That inverse relationship is the whole design: you cannot push intensity and volume up together for long, so the block trades one for the other and the deload week resets fatigue before the test. If you have been stalling on a programme that adds weight and keeps the sets and reps the same, the missing ingredient is usually this trade, not more effort.

Finishing the block does not mean your max went up

This is the point most lifters miss, and it is pure arithmetic. The block above ends with 85kg for 3 reps. Run that through Epley and 85kg × (1 + 3/30) = 93.5kg. Through Brzycki it is 90.0kg. Every prescribed set in the block is submaximal, with about 6 to 7% in hand on the final week, because a block you can only just survive is a block you will fail in week three. So completing it is evidence that you are roughly as strong as the number you programmed from. It is not evidence that you are stronger.

The useful test is what the last set of week three could have been. Take the final set of 3 at 85kg as a set for max reps, stopping one rep short of failure, and read off the estimate:

Reps at 85kg	Epley est.	Brzycki est.	Verdict
3	93.5kg	90.0kg	Repeat the block at 100kg
4	96.3kg	92.7kg	Repeat
5	99.2kg	95.6kg	Repeat, you are at baseline
6	102.0kg	98.7kg	Rebuild from 102.5kg
7	104.8kg	102.0kg	Rebuild from 105kg
8	107.7kg	105.5kg	Your original max was too low

Six reps at 85kg, or equivalently 4 at 90kg (102.0kg by Epley), is the threshold for adding 2.5kg to the next block. Five reps is simply confirmation of the number you already had. This is a cheaper and safer test than a maximal single, and it is what we mean by "if week 3 moved faster than expected, add 2.5 to 5kg and skip the retest". Use the 1RM calculator on the set and let it make the decision.

Programming off a stale max

The opposite error is quieter. If you got stronger and did not update the number, every percentage in your programme is lower than it says. With a true max 5% above the one you are using, a programmed 85% is really 81.0% and a programmed 90% is 85.7%. Let it drift 10% and your "heavy" 90% sets are being done at 81.8%, which is a hypertrophy load, not a strength one. The block still feels like training, but the intensification that was supposed to drive the adaptation has quietly stopped. Retesting or re-estimating every four to six weeks is not about ego, it is about keeping the percentages honest.

How fast progress should actually arrive

Programme-hopping is caused by expectations more than by bad programmes. Broadly, a beginner might add 20 to 40kg to a squat in a first year, an intermediate 10 to 20kg, and an advanced lifter perhaps 5kg across a whole year of well-run training. Convert those into 2.5kg PRs across a year of five-week blocks (about 10.4 blocks) and the picture is sobering:

Training age	Gain per year	Gain per block	2.5kg PRs per year	Time between PRs
Beginner (fast)	40kg	3.85kg	16	3.2 weeks
Beginner (slow)	20kg	1.92kg	8	6.5 weeks
Intermediate (fast)	20kg	1.92kg	8	6.5 weeks
Intermediate (slow)	10kg	0.96kg	4	13 weeks
Advanced	5kg	0.48kg	2	26 weeks

An intermediate gaining 10kg a year is doing fine, and is adding less than one smallest-plate increment per block. A PR every other block is the expected outcome, not a plateau. An advanced lifter making textbook progress sets two PRs a year. If you are in either of those brackets and changing programmes every six weeks because the number did not move, the programme was never given long enough to be judged. Six weeks is barely one block.

Progressive overload without adding weight

Given how small the sustainable weight increments become, most of an intermediate's overload has to come from somewhere other than the bar. In rough order of usefulness: the same weight for an extra rep (worth w/30 of max, as above); an extra set at the working weight (in week one of the block above, a sixth set of 5 at 75kg is 375kg of extra tonnage, a 20% increase); the same session with better bar speed, which is the earliest signal that a load has become submaximal; and the same work in less time. All of these are more total work than last time, and all of them are available on the weeks when 2.5kg is not.

Accessory work should target the sticking point

Accessories should address the position where the lift actually fails. Film the failed rep and watch where the bar slows. If the bench stalls off the chest, paused reps and wider grip work load that position; if it stalls at lockout, close-grip pressing and board work do. If the squat folds forward out of the hole, upper-back and front squat work usually help more than any leg machine. If the deadlift leaves the floor slowly, deficit pulls; if it dies at the knee, blocks or pauses just below it. A generic accessory list trains the whole lift equally, which means most of the work goes to positions that were never the problem.

Recovery is where the number goes up

Training is the stimulus, recovery is the adaptation, and no programme survives sustained neglect of sleep, protein or total stress. Seven to nine hours of sleep; roughly 1.6 to 2.2g of protein per kilogram of bodyweight per day, which is the range most of the strength literature supports; and eating at least at maintenance, because gaining strength in a substantial deficit is possible but slow. If you want a bigger max and are cutting hard at the same time, expect the slow end of every bracket in the progress table.

Test day: what to put on the bar

When you do test a true single, the attempt selection matters more than the warm-up. Three attempts is the standard: an opener you could do for a double on a bad day, a second at or just under your current max, and a third that is the actual PR. Rounded to 2.5kg loading:

Current max	Opener (90%)	Second (97%)	Third (102 to 103%)
100kg	90kg	97.5kg	102.5kg
140kg	125kg	135kg	142.5 or 145kg
180kg	162.5kg	175kg	182.5 or 185kg

And a warm-up ladder that gets you there without spending the max on the way up. Reps fall as the weight rises, so that no single warm-up set is more than a comfortable effort:

Step	Reps	100kg max	140kg max	180kg max
Empty bar	8	20kg	20kg	20kg
40%	5	40kg	55kg	72.5kg
60%	3	60kg	85kg	107.5kg
75%	2	75kg	105kg	135kg
85%	1	85kg	120kg	152.5kg
92%	1	92.5kg	130kg	165kg

Then the opener. Rest three to five minutes between the last warm-up and each attempt. If the opener moves slowly, take the second attempt at your current max and treat it as the day's result rather than forcing the third. A missed third attempt costs more recovery than a made second, and the estimate from your week-three set was probably the better number anyway.

Frequently asked questions

How often should I test my 1RM?

Every four to six weeks at most, and ideally by estimating from a heavy set of three to five reps rather than a true single. A maximal single costs several days of recovery and tells you little that an 85 per cent set for max reps would not. Testing more often than once a block mostly measures fatigue.

Is it better to add weight or add reps?

Below roughly 87kg on the bar, a 2.5kg jump at the same reps raises your estimated max by more than one extra rep at the same weight. Above that, one extra rep is worth more. Under the Epley formula a rep is always worth the weight divided by 30, while 2.5kg is worth 2.5 times (1 + reps/30), so the crossover is at 75 + 2.5 times reps kilograms.

How much should my 1RM go up each year?

Broadly, 20 to 40kg on a squat in a first year of proper training, 10 to 20kg as an intermediate and around 5kg a year once advanced. At 2.5kg increments that is a new PR every three to seven weeks for a beginner, every seven to thirteen weeks for an intermediate and roughly every six months for an advanced lifter.

If I complete every set in my strength block, has my max gone up?

Not necessarily. A block written off a 100kg max ends with 3 sets of 3 at 85kg, and 85kg for 3 only implies a max of about 93.5kg. The sets are submaximal by design. To justify adding 2.5kg to the next block you need something like 85kg for 6 or 90kg for 4, both of which estimate to 102kg.

Method, assumptions and sources

Every figure on this page was computed programmatically from the stated formulas and rounded for display. The rep-versus-plate crossover, w = 75 + 2.5r, follows from setting w/30 equal to 2.5(1 + r/30) and can be checked by hand.
Estimates use Epley, 1RM = w(1 + r/30), and Brzycki, 1RM = w × 36/(37 − r). Both assume a set taken to, or one rep short of, technical failure. Stopping further short understates every estimate here.
The yearly progress brackets (20 to 40kg, 10 to 20kg, ~5kg) are conventional coaching expectations for the squat, not published rates from a single study, and they vary with lift, bodyweight, age and sex. The block-and-PR arithmetic is exact given those inputs; the inputs themselves are assumptions.
The 75/80/85/60% block, the 90/97/102% attempt structure and the warm-up ladder are common conventions, not the only correct ones. The percentages are illustrative; the point is the relationships between them.
Protein and sleep ranges summarise the general direction of the strength and sports nutrition literature rather than a single source, and are not individual dietary advice.
General training information. If you are new to barbell training, returning from injury or have any relevant medical condition, get qualified in-person coaching before working at maximal loads.