← Blog · Method

Progressive overload that actually works

23 July 2026 · 8 min read

Every training article agrees on one sentence: to get stronger, you have to progressively overload. Then almost all of them stop, as if the hard part were the idea rather than the execution. The idea is trivial. Doing it for eleven months in a row, through a bad week at work and a shoulder that complains on Tuesdays, is where programs die.

Here is what progression looks like when you take it seriously.

Adding weight is only one of five levers

Most people equate overload with “more kilos on the bar”. That works for roughly three months, and then the bar stops moving. Load is a product of several variables, and any of them can carry the progression while the others hold still:

• Intensity — more weight for the same reps. The obvious one, and the first to run out.
• Volume — more sets or reps at the same weight. 3×5 at 80 kg becomes 5×5 at 80 kg.
• Density — the same work in less time. Cutting rest from three minutes to two is a real increase in stress.
• Range of motion — a paused rep at the chest is not the same set as a touch-and-go rep.
• Proximity to failure — the same 80 kg taken from three reps in reserve to one is a substantially harder set.

A good program pushes one of these at a time and leaves the rest alone. Pushing three at once is how a training week turns into a stress test with no useful data at the end.

The linear phase is short, and that's fine

If you are new, adding 2.5 kg to the bar every week genuinely works, and you should enjoy it while it lasts. It ends — usually somewhere between the third and the sixth month — and the end is not a failure. It is the point where the method has to change.

Beginners fail because they don't add weight. Intermediates fail because they keep trying to add it the same way.

What replaces it is some form of cycling: a few weeks of accumulating work, then a week where intensity peaks and volume drops, then a deload, then a new cycle starting from a slightly higher base. The bar does not move up every single week any more. It moves up every cycle.

Three ways out of a stall

A stall means the same weight for the same reps in two consecutive attempts, with sleep and food roughly normal. One bad session is not a stall — it is a bad session. When it really is a stall, there are three sane responses.

1. Reset and run it back up

Drop the working weight by about ten percent and rebuild over three to four weeks. You will pass the old sticking point with better technique and more accumulated volume. This is boring and it works, which is why nobody wants to do it.

2. Change the variable

Keep the weight and add a rep. Or keep the weight and reps and add a set. Or pause the rep for a second. The load on the muscle goes up while the number on the bar does not — which also removes the psychological wall that the number itself creates.

3. Attack the weak link, not the lift

Bench failing off the chest is usually not a bench problem — it is a bottom-position strength problem, and paused work or spoto presses fix it. Failing at lockout points at triceps. A stall is information about which part of the movement is behind.

Why a plan has to be rewritten, not followed

Here is the part that a printed program can never do. Progression is a feedback loop: you plan a load, you find out what actually happened, and the next load depends on that result. A twelve-week program printed on day one is a set of guesses about weeks two through twelve — guesses made before a single rep of evidence existed.

Concretely: if your planned top set of 85×5 moved with two reps in reserve, the next jump should be bigger than planned. If it took everything you had and the last rep took four seconds, the next session should repeat 85 rather than march on to 87.5. Same program, opposite decisions, and the only way to know which one applies is to look at what you logged last night.

The short version

• Overload is five variables, not one. Push one at a time.
• Linear progression ends. Plan for cycles before it does.
• Two identical sessions in a row is a stall; one bad session is not.
• Reset ten percent, change the variable, or fix the weak link. Pick one and commit for a month.
• A program that isn't rewritten from your log is a guess with good formatting.

Training information, not medical advice. If something hurts, talk to a professional before you load it.