Progressive Overload: Sets, Reps, Load and RPE

Published October 23, 2026 · Last reviewed October 23, 2026

The variables you can change

Five dials control the stress of a lift: load (the weight), repetitions, sets, effort (how close to failure) and frequency. The 2026 ACSM stand, an overview of 137 systematic reviews covering more than 30,000 participants, reports what each does in healthy adults. For strength, heavier loads (80% of one-repetition maximum or more) were favoured. For power, moderate loads of 30–70% were favoured. For muscle size, load had no consistent effect, and higher weekly volume helped (see the volume guide). Short (under 1 minute) and long (over 1 minute) rests between sets did not change strength, and there was not enough data to say for muscle size. The 2009 stand had suggested 3–5 minutes of rest for heavy strength work and 1–2 minutes for hypertrophy. The newer stand does not give a rest figure, so treat rest as a preference you can adjust to keep your sets honest. For B, C

Two further points from the 2026 stand are worth holding on to. Progression is not required to get benefit but is likely needed by those seeking continued longer-term progress. And periodised programmes were not significantly better than non-periodised ones when progressive overload was applied. So the first job is not a clever calendar. It is a log, a range and a rule for adding.

How to progress, step by step

Action

Use double progression. Choose a range (for example 8–12) and a set count (for example 3). Add repetitions to your weakest set each session. When every set reaches the top, raise the weight by 2–10% and return to the bottom of the range.

Why it helps

The 2009 ACSM stand gave 2–10% as the increase once a lifter can complete one or two repetitions over the target. A range keeps each session a small, achievable step.

Verification

Your log shows the rep count rising within a weight, then a drop-back after each weight increase. Total weight lifted per session trends upward over several weeks. For A, B

Action

Rate every working set with reps in reserve (RIR) or RPE. The RIR-based RPE scale maps RPE 10 to 0 reps in reserve, 9 to 1, 8 to 2, 7 to 3. Aim most sets at RPE 7–8.

Why it helps

ACSM 2026 suggests stopping at about 2–3 in reserve because failure training did not consistently pay off. RPE is a practical way to hold that effort when your weights change from day to day, as the 2016 scale validation reported.

Verification

You write an RPE next to each set and the last set of a session is never an RPE 10 unless you planned a test. The RPE load calculator shows the approximate working weight for your target.

Action

Treat your RIR guess as approximate. The closer a set ends to failure, and the fewer repetitions it takes, the better the guess. Sets of 12 or fewer give the most reliable read.

Why it helps

A 2022 review of 16 publications found lifters underpredicted repetitions to failure by 0.95 repetitions on average (95% CI 0.17 to 1.73), with accuracy better at 12 or fewer repetitions and closer to failure. A 2024 review reached a similar conclusion: RIR estimates are more accurate when made close to failure.

Verification

Once a month, take one safe set closer to failure (with a spotter or in a rack) and check whether your guess matched. Record the gap and adjust your future ratings by it.

Action

Know what RPE can and cannot tell you. A 2022 meta-analysis of 75 studies found an overall validity coefficient of r = 0.88 between RPE and exercise intensity, but heterogeneity was very high, and a 2025 systematic review of 18 studies found moderate correlations between RPE scales, movement velocity and load.

Why it helps

RPE is a useful signal, not a measurement. Sleep, motivation and fatigue shift it, and the 2025 review advises familiarising lifters with the scale and considering factors such as training level and environment.

Verification

If your RPE for the same weight swings by two or more points across a week, treat it as a readiness signal and read the sleep and stress guide before adding weight.

Action

For coaches: For D teach the scale with a set taken close to failure on a safe lift, then ask the lifter to predict her repetitions in reserve before each set and compare.

Why it helps

The 2022 review found accuracy only trivially better in later sets and no meaningful difference by training status, so checking a guess against a real result is the practical way for a lifter to learn her own bias. The 2024 scoping review reports that the most common use of RIR was to set a target and adjust the load to reach it.

Verification

You and the lifter keep a record of predicted and actual repetitions in reserve and look for a consistent gap, such as always one too many.

Double progression in five steps

1. Choose a repetition range for the exercise, such as 8–12, and a number of sets, such as 3.
2. Pick a starting weight that leaves two or three repetitions in reserve on the last set.
3. Add a repetition to your weakest set each session until every set reaches the top of the range.
4. Raise the weight by 2–10% and drop back to the bottom of the range.
5. Log every set so that you compare like with like at the next session.

Where percentages of your maximum fit

You do not need a tested maximum to use percentages. The one-rep max estimator predicts a maximum from a submaximal set, with the caveat that estimates become less reliable at higher repetitions. Use it for planning a heavy block, not as a daily target, and let the effort rating override the percentage when the two disagree. For C Advanced lifters may find the 80%-or-more strength finding relevant when planning a strength-focused phase; the 2026 stand does not set a single prescription.

How it works

Overload works because the body adapts to the stress it has met, so the stress has to rise slowly for adaptation to continue. The rise can come from any dial: one more repetition, a small weight jump, one more set, or the same work done with better control. The 2026 ACSM stand lists increasing relative load (a percentage of your maximum) as one way to support overload.

Double progression is useful because it turns that idea into a rule you can follow without a coach. The repetition range sets a floor and a ceiling. The weight jump is small enough (2–10%, from the 2009 stand) that it usually lands you back inside the range. Effort ratings keep each set honest, and the log makes the trend visible. When the trend stalls for several sessions, go to the deloads and plateaus guide.

The evidence has limits. The 2026 stand states that it could not quantify exact RIR and RPE targets, and many underlying trials are small. Use the ranges as starting points and your own log as the test.

Worked example

Suppose, with hypothetical numbers, you press 12 kg dumbbells for 3 sets with a target of 8–12 repetitions.

• Week 1: 8, 8, 8 repetitions. Total repetitions: 24. Volume load: 24 × 12 = 288 kg.
• Week 2: 10, 9, 8. Total: 27. Volume load: 27 × 12 = 324 kg.
• Week 3: 12, 11, 10. Total: 33. Volume load: 33 × 12 = 396 kg.
• Week 4: 12, 12, 12. Total: 36. Volume load: 36 × 12 = 432 kg. Every set is at the top of the range.

Now raise the weight. 2–10% of 12 kg is 0.24–1.2 kg, so the new weight is between 12.2 and 13.2 kg. A 13 kg dumbbell is 13 ÷ 12 = 1.083, an 8.3% increase, inside the range. Week 5 drops back to 8, 8, 8 at 13 kg: 24 × 13 = 312 kg. Compare it with week 1 (288 kg): 312 − 288 = 24 kg more, or 24 ÷ 288 = 8.3%. Compare it with week 4 (432 kg): it is lower, which is expected right after a weight jump. Judge progress by comparing the same point in each cycle, such as week 1 with week 5, not the first week after a jump with the last week before it.