Progressive Overload, in Numbers
Progressive overload is the one principle nearly everyone agrees on: to keep adapting, the demand has to keep rising. Where it gets vague is what exactly should rise, by how much, and how you would know.
Usefully, all the common answers are measurable.
The four things that can go up
- Weight. Same sets and reps, heavier bar. The most direct form, and the first to run out.
- Reps. Same weight, more reps per set. 3 × 8 at 100kg becoming 3 × 10 is a 25% increase in work.
- Sets. Same weight and reps, more of them.
- Density. Same work, less rest. Harder to log honestly and easy to overrate.
The first three all show up in tonnage, which is why sets × reps × weight is the standard proxy. The fourth does not, which is one of tonnage's blind spots.
Why weight alone stops working
A beginner adds 2.5kg to their squat every session and it works for months. Then it stops, and the usual response — try harder, add it anyway — produces a failed set and a lost week.
Look at the percentages and the reason is obvious. On a 100kg max, adding 2.5kg is a 2.5% jump. On a 180kg max, the same 2.5kg is 1.4%. But the absolute jumps that are available to you do not shrink as you get stronger — the plates are the same size. So each increment is a smaller relative step, while the recovery cost of each session keeps rising.
What replaces it is rep progression and set progression, which offer much finer control. Going from 3 × 5 to 3 × 6 at 140kg adds 420kg of tonnage without touching the bar — a real increase in demand that no plate on the rack could deliver as precisely.
Making it concrete
A worked week from the Training Volume Calculator: three squat sessions, plus bench, press and deadlift, totalling 12,050kg. Suppose next week you add one set of five at 120kg to Monday. That is 600kg, or a 5% weekly increase — substantial, and about as much as most people absorb week on week.
Doing that for eight weeks straight would put you at roughly 17,800kg, which almost nobody recovers from. Which is the point of the numbers: they make it obvious when a plan is arithmetically unsustainable long before your body tells you.
The rate that actually works
There is no universal figure, but some grounded reference points:
- Novices can often add weight session to session for months. Take it while it lasts; it does not come back.
- Intermediates generally progress across a block rather than a session — a few percent on the main lifts over four to six weeks.
- Advanced lifters may add a few percent a year, which is why their programming looks so much more elaborate. Small margins need careful management.
The mistake is applying a novice rate to an intermediate body. Adding 2.5kg a week to a squat for a year implies 130kg of annual progress. Nobody does this, and a plan that assumes it will fail in month three — not because of insufficient effort, but because the arithmetic was never possible.
Overload is not monotonic
The word "progressive" misleads people into thinking every week must exceed the last. In practice, load accumulates faster than recovery, and periodic reductions are what let the trend keep rising.
A block that climbs for three weeks and drops 40% in the fourth will out-produce one that grinds upward until it breaks. Your tonnage log should look like a rising sawtooth, not a straight line — and a deload week showing up as a 40% fall is the plan working, not a bad week.
What to actually do
Pick one variable to progress at a time and hold the others still, or you will not know what caused what. Log the working weights, the sets and the reps — three numbers, ten seconds. Check the percentages occasionally to confirm the intensity is where you think it is, and glance at the tonnage monthly rather than daily.
That is enough structure to tell progress from drift, which is really all the measurement is for.
Plateaus that are not plateaus
Before restructuring a programme because progress stopped, it is worth ruling out the things that look like a plateau and are not.
Two bad weeks. Normal variation. Performance moves with sleep, food, stress and time of day, and a fortnight of flat numbers is well inside that noise. Give it another two weeks before concluding anything.
An estimate that was optimistic. If the whole block was built on a max estimated from a set with generous rep counting, every working weight has been slightly too heavy from day one. This does not present as obvious failure — it presents as a programme that was always slightly too hard. Re-estimate honestly with the 1RM calculator.
Reps quietly slipping. The most common self-deception. The bar weight has gone up for six weeks and the sets have gone from a clean five to a scrappy four. Total work is flat or falling while the log looks like progress. This is why reps belong in the log next to the weight.
Missing recovery, not stimulus. If sleep has collapsed or you are in a deficit, adding training volume is the wrong lever, and pulling it may be the right one.
Autoregulation: letting the day decide
Fixed percentages assume every session finds you in the same state. They do not. Autoregulation adjusts the day's work to how the day is actually going.
The simplest usable version needs no equipment: a top set taken to a target effort level, then back-off sets computed from what that top set actually was rather than from a number decided three weeks ago. On a good day the weights are higher; on a bad day the session still happens at an appropriate load instead of turning into three failed attempts.
A second version uses reps in reserve — how many more you could have done — as the prescription. "Three sets of five leaving two in the tank" adapts itself; "three sets of five at 85%" does not.
Neither replaces having a plan. They replace the assumption that the plan knows more about today than you do. Percentages from the percentage calculator still set the range you are working in — autoregulation just decides where inside it you land.
Why the log is the whole mechanism
Everything in this article depends on a record. Without one you cannot tell a plateau from a bad fortnight, cannot see reps slipping under rising weight, and cannot know whether your weekly work is up or down.
Three numbers per set — weight, sets, reps — and a note when something unusual happened. That is enough to compute your tonnage, estimate your max, check your intensities, and answer every question above. It takes seconds and it is the difference between training and merely exercising with intent.