Progressive Overload Explained: Science of Building Strength & Muscle
Progressive overload means gradually increasing the demand you place on a muscle so it has a reason to keep adapting. The practical answer to how much weight to add is smaller than most programs imply: a trained lifter gains roughly 1 to 2 kg a month on a squat or deadlift and about half that on a bench press, so the useful rule is not a weekly number but a trigger, add load only when you have earned it by hitting the top of your rep range.
There is direct evidence that the progression itself matters. Kassiano and colleagues randomised the arms of 55 untrained young women to progressive overload or a fixed load for eight weeks of elbow extension, three days a week (Medicine and Science in Sports and Exercise, 2026, volume 58, pages 1556 to 1565). Summed triceps thickness rose 22.9% in the progressive condition, 11.6% in the fixed-load condition and 1.8% in non-exercising controls, P less than 0.001. Training without progression still built muscle. It built about half as much.
The six levers of progression, with worked examples
Lever | Before | After | When this is the right lever to pull |
|---|---|---|---|
Add load | Bench press 3 sets of 8 at 60 kg, last set felt like 1 rep left | Bench press 3 sets of 8 at 62.5 kg | You hit the top of your rep range on every set with a repetition or two still in reserve. This is the default lever and the best supported one |
Add repetitions | Dumbbell row 3 sets of 8 at 30 kg | Dumbbell row 3 sets of 11 at 30 kg | The next available jump is too big. Dumbbells often step 2.5 kg per hand, which can be a 10% increase on an upper body lift. Add reps until the jump becomes affordable |
Add sets | Leg press 3 sets of 10 at 140 kg, 6 direct quadriceps sets a week | Leg press 4 sets of 10 at 140 kg, 8 direct sets a week | You are below about 10 hard sets per muscle group per week and recovering easily. This is the lever for size more than for maximal strength |
Add frequency | Squat once a week, 5 sets of 5 at 100 kg | Squat twice a week, 3 sets of 5 at 100 kg each session, 6 sets total | Your last two sets in a session have stopped resembling the first two. Splitting the work restores quality and quietly adds a set |
Reduce rest | Leg extension 3 sets of 12 at 45 kg, 3 minutes rest | Same sets and reps, 90 seconds rest | Rarely. Use it only when time is the binding constraint, and only if reps and load hold. Below about 90 seconds it starts to cost you load rather than add stimulus |
Improve range of motion or control | Back squat 3 sets of 5 at 105 kg to a shallow depth, bar speeding up and reps shortening | Back squat 3 sets of 5 at 90 kg to full depth with a controlled descent, rebuilding from there | Your reps have been getting shorter and faster as the load has climbed. This is the honest lever when the bar weight is rising but the work is not |
Sources and derivation: the load and repetition rows follow Kassiano and colleagues (Medicine and Science in Sports and Exercise, 2026) and Plotkin and colleagues (PeerJ, 2022, volume 10, article e14142). The volume, load and frequency thresholds follow the American College of Sports Medicine 2026 Position Stand, which synthesised 137 systematic reviews and more than 30,000 participants (Medicine and Science in Sports and Exercise, 2026, volume 58, pages 851 to 872). The rest interval row follows Singer and colleagues (Frontiers in Sports and Active Living, 2024, volume 6, article 1429789) and Grgic and colleagues (Sports Medicine, 2018, volume 48, pages 137 to 151). The specific weights and repetitions are constructed examples, not measured data.
Adding load and adding reps work about equally well
Plotkin and colleagues randomised 43 people with at least a year of consistent lower body training to either increase load while holding repetitions constant, or increase repetitions while holding load constant, twice a week for eight weeks. Rectus femoris growth modestly favoured the repetition group, 2.8 mm summed across sites (90% confidence interval minus 0.5 to 5.8). Dynamic strength slightly favoured the load group, 2.0 kg (90% confidence interval minus 2.4 to 7.8). Every other outcome sat near zero: muscle thicknesses under 1 mm apart, endurance under 1%, jump height 0.1 cm.
So the lever matters less than the fact that you are pulling one. Load progression suits strength slightly better, repetition progression suits size slightly better, and both differences are small enough that convenience should decide. That matters because the smallest plate in most gyms is 1.25 kg per side. On a 40 kg bench press that is a 6% jump, which is a lot to ask week after week.
How much you can realistically expect to add
Stage | Squat | Bench press | Deadlift | What the number actually measures |
|---|---|---|---|---|
First 5 to 6 months, previously untrained | About 21% strength gain over the whole period, roughly 3.5 to 4.5% a month, with enormous spread | 287 untrained adults aged 19 to 78, 20 to 24 weeks of periodised training twice a week. Mean strength change 21.1% plus or minus 11.5%, individual range minus 8% to plus 60% (Ahtiainen and colleagues, Age, 2016) | ||
First year, low-dose training | Roughly 30 to 50% over the year, with the curve flattening by 12 to 24 months | 14,690 people training once a week, single sets to failure on six exercises, tracked up to 6.8 years. Strength reached about 50 to 60% above baseline by year six (Steele and colleagues, Research Quarterly for Exercise and Sport, 2023) | ||
Competitive lifters, best rate | 1.7 to 2.1 kg per month | 0.9 to 1.1 kg per month | 1.5 to 1.8 kg per month | 1,896 powerlifters across 407 meets, 2003 to 2018. Maximum annual adaptation: squat 20.2 to 25.4 kg, deadlift 18.1 to 21.1 kg, bench press 10.5 to 12.8 kg, P less than or equal to 0.001 (Latella and colleagues, Medicine and Science in Sports and Exercise, 2022) |
Competitive lifters, typical rate across a career | About 1.2 to 1.5 kg per month per lift | Same cohort. Overall gain of 0.15 kg per day for men and 0.12 kg per day for women on the three-lift total, which is 44 to 55 kg a year spread across three lifts (Latella and colleagues, Journal of Strength and Conditioning Research, 2020) | ||
Ten years in | About 7.5 to 12.5% in the first year of competing, reaching about 20% total after ten years | Growth modelling of powerlifting competition results by age category. The strongest quartile of men gained at 0.102 kg per day against 0.211 kg per day for the least strong, P equals 0.010 (Latella and colleagues, Sports Medicine, 2024, and Journal of Strength and Conditioning Research, 2020) | ||
Sources and limits: the per-lift kilogram figures are the published annual rates divided by twelve, which assumes an even spread across the year that real training does not have. The novice rows are percentages because no study reports novice progression in kilograms per lift, so treat any kilogram figure you derive from them as arithmetic rather than measurement. The powerlifting data come from competition results, which means the athletes were peaking, cutting to a weight class and occasionally injured, and the samples are self-selected competitors rather than the general population.
Three things fall out of that table. Upper body lifts progress at roughly half the rate of lower body lifts, which is not a personal failing but a consistent finding across nearly 1,900 athletes. The rate declines steeply with training age. And the individual spread is very wide: Ahtiainen's 287 participants ranged from minus 8% to plus 60% on the same program.
Double progression, and how to run it
Double progression means you progress two variables in sequence rather than one continuously. Pick a rep range, say 3 sets of 6 to 9. Keep the load fixed until you hit the top of the range on all sets with a repetition or two left. Then add the smallest available increment and drop back to the bottom of the range.
A worked cycle on the bench press
Training twice a week, with the load held until the top of the range is reached on all three sets:
Week 1: 60 kg for 7, 6, 6.
Week 2: 60 kg for 8, 7, 7.
Week 3: 60 kg for 9, 8, 8.
Week 4: 60 kg for 9, 9, 9. Range topped out on all three sets.
Week 5: 62.5 kg for 7, 6, 6. Cycle restarts.
That is 2.5 kg in five weeks, about 0.5 kg a week, which lands close to the 0.9 to 1.1 kg a month that competitive lifters actually achieve on the bench press. No published trial has tested double progression directly against a fixed weekly increment, so it is worth being clear that this is a sensible way of combining two supported ingredients rather than a proven superior method. Those ingredients are that progression matters (Kassiano) and that adding reps and adding load produce similar adaptations (Plotkin).
Autoregulation against fixed percentages
Autoregulation adjusts the load to how you are performing today, either by feel using repetitions in reserve, or objectively using bar velocity. Fixed percentage training sets the load from a percentage of your one repetition maximum regardless of the day.
Hickmott and colleagues meta-analysed 15 studies in resistance-trained people (Sports Medicine Open, 2022, volume 8, article 9). Comparing autoregulated load prescription against percentage-based prescription, there was no significant difference in one repetition maximum strength, mean difference 2.07 kg (95% confidence interval minus 0.32 to 4.46, p equals 0.09). Volume autoregulation did produce a signal: velocity loss thresholds at or below 25% gave more strength, 2.32 kg (95% confidence interval 0.33 to 4.31, p equals 0.02), and less growth, 0.61 cm² (95% confidence interval 0.05 to 1.16, p equals 0.03), than thresholds above 25%. Stopping sets earlier favours strength, pushing them further favours size.
Wang and colleagues reached the same conclusion for maximal strength across 17 studies and 348 trained participants (BMC Sports Science, Medicine and Rehabilitation, 2026, volume 18, article 57). Velocity-based training beat percentage-based training for jump height, standardised mean difference 0.27 (95% confidence interval 0.03 to 0.51), and change of direction, 0.45 (0.17 to 0.73), but not for maximal strength, 0.21 (minus 0.01 to 0.43, p equals 0.064).
One analysis disagrees. Huang and colleagues ran a network meta-analysis and ranked autoregulating progressive resistance exercise ahead of percentage-based training on both squat and bench press (Journal of Exercise Science and Fitness, 2025, volume 23, pages 360 to 369). Network rankings are sensitive to sparse indirect comparisons, and Hickmott's direct pairwise analysis is the more conservative read, so that is the one used here. The practical point survives either way: autoregulation is a way to keep the load honest on a bad week, not a faster route.
Whether you need to train to failure
You do not. Grgic and colleagues meta-analysed 15 studies and found no significant difference between training to failure and stopping short of it, for strength (effect size minus 0.09, 95% confidence interval minus 0.22 to 0.05) or for muscle size (effect size 0.22, 95% confidence interval minus 0.11 to 0.55) (Journal of Sport and Health Science, 2022, volume 11, pages 202 to 211). Where volume was not equated, non-failure training actually favoured strength, effect size minus 0.32 (95% confidence interval minus 0.57 to minus 0.07). In resistance-trained participants specifically, failure showed a small advantage for size, effect size 0.15 (0.03 to 0.26).
Refalo and colleagues found the same shape across 15 studies (Sports Medicine, 2023, volume 53, pages 649 to 665). Momentary muscular failure against non-failure gave an effect size of 0.12 for hypertrophy (95% confidence interval minus 0.13 to 0.37, p equals 0.343). The 2026 ACSM Position Stand, working from a much larger evidence base, concluded that training to momentary muscle fatigue did not consistently affect outcomes.
The trade-off is fatigue. Sets taken to failure cost more recovery for a similar stimulus, and that cost tends to show up as a stalled progression two sessions later rather than on the day. Leaving one or two repetitions in reserve on most sets, and saving failure for the last set of an isolation exercise, keeps the stimulus and returns the recovery.
Deloads and periodisation
The deload evidence is thinner than its popularity suggests. Coleman and colleagues randomised 39 trained lifters to a one-week break at the midpoint of a nine-week program or to continuous training (PeerJ, 2024, volume 12, article e16777). Muscle size, power and endurance did not differ. Continuous training produced greater isometric and dynamic strength gains. Pancar and colleagues used a within-subject design in 19 untrained men, reducing volume and frequency in weeks 4 and 8, and found no interaction for muscle thickness or strength endurance, with all confidence intervals for the difference including zero (Scientific Reports, 2026, volume 16, article 10299).
Neither study supports a scheduled deload as a growth strategy. Both are consistent with deloading as a fatigue management tool, used when performance is falling rather than on a calendar. Both are also short, and neither followed lifters long enough to see the accumulated fatigue that deloads are meant to address.
Periodisation is similar. Moesgaard and colleagues meta-analysed 35 volume-equated studies and found periodised training beat non-periodised for one repetition maximum strength, effect size 0.31 (95% confidence interval 0.04 to 0.57, p equals 0.02), with no difference for muscle size, effect size 0.13 (minus 0.10 to 0.36) (Sports Medicine, 2022, volume 52, pages 1647 to 1666). Undulating periodisation beat linear for strength in trained participants, effect size 0.61, but not in untrained ones. The 2026 ACSM Position Stand concluded periodisation did not consistently affect outcomes. Read together: a small strength benefit for trained lifters, and nothing worth reorganising a beginner's program around.
What actually causes a plateau
The most useful data here come from Steele and colleagues, who modelled strength records from 14,690 people, 60% female, average age 48, training once a week with single sets to failure on six exercises, for up to 352 weeks (Research Quarterly for Exercise and Sport, 2023, volume 94, pages 913 to 930). Strength rose about 30 to 50% in the first year and then flattened asymptotically, reaching only about 50 to 60% above baseline six years later. Sex, bodyweight and age barely moved the curve.
That is the honest picture. A plateau is mostly the normal shape of the adaptation curve, not a mistake you made. Latella's powerlifting data show the same thing from the other end, with the strongest quartile of men gaining at less than half the rate of the least strong, which the authors read as a ceiling effect.
What to check when you stall
Steele's group were explicit that they could not tell whether the plateau can be overcome by changing the approach. So what follows is inference from the dose-response evidence rather than a tested fix. If you have stalled, check the dose against the thresholds the ACSM synthesis identified: at least 10 hard sets per muscle group per week for size, loads at or above 80% of your one repetition maximum for strength, a complete range of motion, at least two sessions a week, and the lift you care about placed early in the session. Rhea and colleagues, meta-analysing 140 studies and 1,433 effect sizes, found untrained people gain most at around 60% of one repetition maximum while trained people need around 80% (Medicine and Science in Sports and Exercise, 2003, volume 35, pages 456 to 464). A load that worked in year one is often simply too light in year three.
Then change the lever instead of forcing the same one. If the bar will not move, add a set, add a session, restore the range of motion you have been quietly trimming, or hold the weight and chase reps for a month.
What a month of progression looks like
A constructed example for an intermediate lifter training four days a week, running double progression on the main lifts with 3 sets of 5 to 8 repetitions, stopping one or two repetitions short of failure.
Week | Squat | Bench press | Romanian deadlift | Chest-supported row |
|---|---|---|---|---|
1 | 100 kg for 6, 5, 5 | 70 kg for 6, 6, 5 | 90 kg for 8, 8, 7 | 60 kg for 10, 10, 9 |
2 | 100 kg for 7, 6, 6 | 70 kg for 7, 6, 6 | 90 kg for 8, 8, 8, load goes up next week | 60 kg for 10, 10, 10, load goes up next week |
3 | 100 kg for 8, 7, 7 | 70 kg for 7, 7, 6 | 95 kg for 8, 7, 7 | 65 kg for 9, 8, 8 |
4 | 100 kg for 8, 8, 8, load goes up next week | 70 kg for 8, 7, 7 | 95 kg for 8, 8, 7 | 65 kg for 10, 9, 9 |
5 | 102.5 kg for 6, 5, 5 | 70 kg for 8, 8, 7 | 95 kg for 8, 8, 8, load goes up next week | 65 kg for 10, 10, 10, load goes up next week |
Source: constructed example. The progression rate is set to match the published rates in the second table rather than invented.
Over five weeks the squat moved 2.5 kg and the bench press moved nothing at all, having gone from 17 total repetitions to 23 at the same load. That is what real progression looks like once you are past the first year, and it still adds up to roughly 25 kg on the squat across a year, which is the top of the range competitive powerlifters achieve.
Where your own data fits
Every rate above is a population average sitting on top of a very wide individual spread, and the largest single input into whether this week's session goes up or sideways is how well you have recovered from last week's. Your sleep, resting heart rate and heart rate variability, and your session-to-session bar performance are all measurable rather than guessable. hlth.coach reads your wearable and body composition data and adjusts training and recovery against what your body is doing that week, which is a better basis for deciding whether to add the 2.5 kg than a number written into a spreadsheet a month ago.
The short version
Progression itself matters. Fixed-load training built about half the muscle of progressive training over eight weeks in 55 untrained women.
Adding load and adding repetitions work about equally well. Load favours strength slightly, reps favour size slightly, and both differences are small.
Realistic rates for a trained lifter are 1.7 to 2.1 kg a month on the squat, 1.5 to 1.8 on the deadlift and 0.9 to 1.1 on the bench press, from nearly 1,900 competitive powerlifters.
Untrained people gain much faster, about 21% strength in 20 to 24 weeks, with individual results ranging from minus 8% to plus 60%.
Double progression is the practical way to run it: fix the load, climb the rep range, then add the smallest available increment and start again.
Autoregulation matches percentage-based loading for maximal strength. Its value is protecting bad weeks, not accelerating good ones.
Training to failure is not required for strength or size, and it costs recovery that your next session needs.
Scheduled deloads did not improve outcomes in either trial that tested them, and one favoured continuous training for strength. Deload when performance drops, not on a calendar.
Plateaus are mostly the normal shape of the curve. Strength flattened by 12 to 24 months in 14,690 low-dose trainees. Check your dose against the thresholds, then change the lever.
Common questions
How much weight should I add each week?
Small increments, and less than you think. A common rule of thumb caps progression at 10 percent per week, but that is impractical at both ends: a beginner on light loads waits too long, and an experienced lifter cannot add 10 percent to a heavy squat. The table above gives realistic rates by lift and training age.
Should I increase reps or weight?
Both work. A controlled comparison found load progression and repetition progression produced similar muscular adaptations, with load slightly favouring strength and reps slightly favouring size. The lever matters less than the fact that something is progressing.
What is the 2 for 2 rule?
It says add weight once you can complete two extra reps beyond your target on the last set, in two consecutive sessions. It is a reasonable default because it makes the decision automatic rather than a judgement call in the moment.
Do I need to progressively overload every week?
No, and expecting to is how people end up ego-lifting or stalling. Weekly progression works while increments stay small. Beyond the early phase, progress arrives in steps with maintenance weeks between.
Do I have to train to failure to progress?
No. Progression needs the stimulus to increase over time, not for every set to end in failure. Training close to failure gives most of the benefit at a fraction of the fatigue cost.
Why have I stopped progressing?
Usually recovery, not effort. Sleep, food and accumulated fatigue limit progress long before willpower does. Before adding more work, check whether you are recovering from what you already do, and consider a lighter week.
Sources
Kassiano W, Santos-Melo V, Manske I, Lisboa F, Miguel A, Gomes F, Prado A, Stavinski N, Costa B, Cyrino ES. Progressive overload affects the magnitude of muscle hypertrophy. Medicine and Science in Sports and Exercise, 2026, 58(7), 1556 to 1565.
Plotkin D, Coleman M, Van Every D, Maldonado J, Oberlin D, Israetel M, Feather J, Alto A, Vigotsky AD, Schoenfeld BJ. Progressive overload without progressing load? The effects of load or repetition progression on muscular adaptations. PeerJ, 2022, 10, e14142.
Currier BS, D'Souza AC, Fiatarone Singh MA, Lowisz CV, Rawson ES, Schoenfeld BJ, Smith-Ryan AE, Steen JP, Thomas GA, Triplett NT, Washington TA, Werner TJ, Phillips SM. American College of Sports Medicine Position Stand. Resistance training prescription for muscle function, hypertrophy, and physical performance in healthy adults: an overview of reviews. Medicine and Science in Sports and Exercise, 2026, 58(4), 851 to 872.
Ahtiainen JP, Walker S, Peltonen H, Holviala J, Sillanpaa E, Karavirta L, Sallinen J, Mikkola J, Valkeinen H, Mero A, Hulmi JJ, Hakkinen K. Heterogeneity in resistance training-induced muscle strength and mass responses in men and women of different ages. Age, 2016, 38(1), 10.
Steele J, Fisher JP, Giessing J, Androulakis-Korakakis P, Wolf M, Kroeske B, Reuters R. Long-term time-course of strength adaptation to minimal dose resistance training through retrospective longitudinal growth modeling. Research Quarterly for Exercise and Sport, 2023, 94(4), 913 to 930.
Latella C, Owen PJ, Davies T, Spathis J, Mallard A, van den Hoek D. Long-term adaptations in the squat, bench press, and deadlift: assessing strength gain in powerlifting athletes. Medicine and Science in Sports and Exercise, 2022, 54(5), 841 to 850.
Latella C, Teo WP, Spathis J, van den Hoek D. Long-term strength adaptation: a 15-year analysis of powerlifting athletes. Journal of Strength and Conditioning Research, 2020, 34(9), 2412 to 2418.
Latella C, van den Hoek D, Wolf M, Androulakis-Korakakis P, Fisher JP, Steele J. Using powerlifting athletes to determine strength adaptations across ages in males and females: a longitudinal growth modelling approach. Sports Medicine, 2024, 54(3), 753 to 774.
Hickmott LM, Chilibeck PD, Shaw KA, Butcher SJ. The effect of load and volume autoregulation on muscular strength and hypertrophy: a systematic review and meta-analysis. Sports Medicine Open, 2022, 8(1), 9.
Wang Y, Qiu J, Dai D, Li Q, Wang X, Luo Q. The effects of velocity-based vs. percentage-based resistance training on sports performance in trained individuals: a systematic review and meta-analysis. BMC Sports Science, Medicine and Rehabilitation, 2026, 18(1), 57.
Huang Z, Sun J, Li D, Chen C, Wang D. Autoregulated resistance training for maximal strength enhancement: a systematic review and network meta-analysis. Journal of Exercise Science and Fitness, 2025, 23(4), 360 to 369.
Grgic J, Schoenfeld BJ, Orazem J, Sabol F. Effects of resistance training performed to repetition failure or non-failure on muscular strength and hypertrophy: a systematic review and meta-analysis. Journal of Sport and Health Science, 2022, 11(2), 202 to 211.
Refalo MC, Helms ER, Trexler ET, Hamilton DL, Fyfe JJ. Influence of resistance training proximity-to-failure on skeletal muscle hypertrophy: a systematic review with meta-analysis. Sports Medicine, 2023, 53(3), 649 to 665.
Coleman M, Burke R, Augustin F, Pinero A, Maldonado J, Fisher JP, Israetel M, Androulakis Korakakis P, Swinton P, Oberlin D, Schoenfeld BJ. Gaining more from doing less? The effects of a one-week deload period during supervised resistance training on muscular adaptations. PeerJ, 2024, 12, e16777.
Pancar Z, Ilhan MT, Darendeli MK, Karaca B, Ikidag MA, Tasdogan AM, Ulema MS, Alkhamees NH, Al-Mhanna SB, Batrakoulis A. Effects of deload periods in resistance training on muscle hypertrophy and strength endurance in untrained young men using a randomized within subject design. Scientific Reports, 2026, 16(1), 10299.
Moesgaard L, Beck MM, Christiansen L, Aagaard P, Lundbye-Jensen J. Effects of periodization on strength and muscle hypertrophy in volume-equated resistance training programs: a systematic review and meta-analysis. Sports Medicine, 2022, 52(7), 1647 to 1666.
Rhea MR, Alvar BA, Burkett LN, Ball SD. A meta-analysis to determine the dose response for strength development. Medicine and Science in Sports and Exercise, 2003, 35(3), 456 to 464.
Singer A, Wolf M, Generoso L, Arias E, Delcastillo K, Echevarria E, Martinez A, Androulakis Korakakis P, Refalo MC, Swinton PA, Schoenfeld BJ. Give it a rest: a systematic review with Bayesian meta-analysis on the effect of inter-set rest interval duration on muscle hypertrophy. Frontiers in Sports and Active Living, 2024, 6, 1429789.
Grgic J, Schoenfeld BJ, Skrepnik M, Davies TB, Mikulic P. Effects of rest interval duration in resistance training on measures of muscular strength: a systematic review. Sports Medicine, 2018, 48(1), 137 to 151.
Kassiano W, Costa B, Nunes JP, Ribeiro AS, Schoenfeld BJ, Cyrino ES. Which ROMs lead to Rome? A systematic review of the effects of range of motion on muscle hypertrophy. Journal of Strength and Conditioning Research, 2023, 37(5), 1135 to 1144.
This article is general information and is not medical advice. Speak with your doctor or an accredited exercise physiologist before starting a new training program, particularly if you have an existing health condition or injury.