A mesocycle is a way of grouping training weeks. The version you have probably read (four to six weeks of rising volume, then one week at half volume) comes from a coaching model, and the specific claim has never been tested. What has been measured is how much volume you do per week. This article separates the two.
The word simply means a training block of several weeks that is planned as a unit. That much is uncontroversial and useful: it is easier to think in blocks than in single sessions.
Everything beyond that (how long the block should be, whether volume should climb inside it, and what the week at the end should look like) is where the evidence and the popular advice part ways.
The widely repeated version runs like this: weeks 1–2 at MEV (minimum effective volume), weeks 3–4 at MAV (maximum adaptive volume), weeks 5–6 approaching MRV (maximum recoverable volume), then a deload week at half volume. Presented that way it sounds like a finding. It is a framework.
The model comes from the Renaissance Periodization circle. Its only peer-reviewed form is a review article with no measurement of its own: Israetel M, Feather J, Faleiro TV, Juneau CE, "Mesocycle Progression in Hypertrophy: Volume Versus Intensity", Strength and Conditioning Journal 2020;42(5):2–6. The abstract says outright that the recommendation is drawn from existing literature: the likely answer, it says, is "some of all 3".
Two details are worth keeping. First, the article itself says "typically 4–8 weeks" before a recovery week, not the 4–6 that circulates. Second, a letter to the editor in the same journal noted that the recommendations are "inferred", and named the gap precisely:
"To our knowledge, no studies have matched total volume while comparing a group increasing weekly sets with a group that maintains a fixed number of sets week to week."
Minor B, Helms E, Schepis J, Letter to the Editor, Strength and Conditioning Journal 2020
Search for a source on the 4-to-6-week cycle with a halved deload and you will land on Painter 2012. It does not support it. That trial ran ten weeks straight, the block group trained with roughly 35 % less volume, the strength result was null, and there was no hypertrophy measure at all.
Practice does not supply the number either. A survey of 246 strength and physique athletes found deloads every 5.6 ± 2.3 weeks lasting 6.4 ± 1.7 days, and the volume reduction coaches report most often is 0–25 %, not 50. Both describe what people do. Neither measures what it achieves.
On the 50 % figure itself: it is unsupported, not refuted. The full text of the review it is usually attributed to sits behind a paywall; its abstract contains no percentage, but a seventy-word abstract cannot rule out a recommendation buried in five printed pages.
In 2026 the American College of Sports Medicine replaced its 2009 position stand with an overview of reviews. For hypertrophy it names two things that helped: higher weekly volume and eccentric overload, the latter meaning supramaximal loading in the lowering phase, not simply lowering the weight slowly. Frequency and load affect strength, not muscle size.
Periodization sits on the other side of that list, among the variables for which no influence on hypertrophy was shown, across four reviews totalling 1,414 participants. The word deload does not appear anywhere in the full text.
"No influence shown" is not the same as "no influence". It means anyone recommending the structure cannot point to this body of evidence for it.
From the position stand: hypertrophy was enhanced by higher volumes, ≥ 10 sets per muscle group per week. The same paper puts an upper bound on it: diminishing returns beyond roughly 18–20 weekly sets, with the dose-response curve described as flattening.
A second number sounds almost identical and means something else. One group built a genuine six-week ramp, from 10 to 32 sets, and tracked it with muscle biopsies. Across the whole block there was no net fibre growth, and the authors describe roughly 20 sets per exercise per week as a volume beyond which little further adaptation occurs.
| Number | Unit | Source |
|---|---|---|
| ≥ 10 sets | per muscle group per week | ACSM position stand 2026: lower bound |
| ~ 18–20 sets | per muscle group per week | ACSM position stand 2026: diminishing returns above this |
| ~ 20 sets | per exercise per week | Vann/Haun: from a trial whose main result was null |
The last two are nearly the same figure and cannot be converted into each other. Bench press, incline press and flyes are three exercises for one muscle group. A ceiling "per exercise" is a different quantity from one "per muscle group", and the second comes from a trial whose headline finding was an absence.
Ten only means something if you know what was counted. Do warm-up sets count? Does bench press count for triceps: fully, half, or not at all? The common answer is: directly worked muscles count fully, assisting muscles count half. That is an agreed way of counting, not a measurement. Before comparing two volume figures, check that both use the same one.
The practical core of the model is to add sets week by week. That has now been tested four times, randomised, against a constant weekly volume. In none of the four did the group that increased its volume grow measurably more.
That removes the justification for staircasing weeks one through six. It does not show the practice is harmful. It shows the advantage is unsupported.
Two randomised trials have set a break week against continuous training and measured muscle thickness by blinded ultrasound. Neither found a hypertrophy advantage. In one of them, training straight through came out ahead on strength, and Mike Israetel, whose model this is, is a co-author of that paper.
Neither tests the prescription in question. One removed the training week entirely rather than halving the volume; the other used a three-week loading block at constant volume. That difference cuts both ways: it means the specific 50 % week has not been shown to work, and it has not been shown to fail either.
The closest thing to a direct test of the whole structure is the six-week ramp mentioned above, which ended in a recovery week and took muscle biopsies throughout. Fibre cross-sectional area at the end of the block was not different from the start.
It cannot settle the deload question, because both of its arms took a lighter week: there was no group that trained straight through. What it does show is that a full ramp of that shape, in that population, produced no net fibre growth over six weeks.
Count your working sets per muscle group per week. Get to at least ten. Stay below roughly twenty, where the added benefit flattens out. Beyond that, the block structure is yours to keep or drop, but keep it because you like planning in blocks, not because the evidence requires it.
If you take a lighter week, nothing in this evidence says it costs you muscle. Nothing says it gains you any either. That is a smaller claim than the model makes, and it is the one that holds.
No, it is untested, which is a different thing. What has been tested, and showed no advantage, is the week-to-week volume increase (four randomised trials) and the hypertrophy benefit of a break week (two trials). The terms themselves have never been set against an alternative with muscle size as the outcome.
There is no measured answer. The review most often cited says "typically 4–8 weeks", and says so as a recommendation drawn from other literature, not as a finding. Athletes surveyed report deloading every 5.6 ± 2.3 weeks. That is a description of behaviour.
Per muscle group per week, as stated in the position stand. The other figure in this article, around twenty, is per exercise, comes from a different paper, and does not convert into the first.
The app shows your weekly volume across the seven muscle groups, the same unit the position stand uses. The AI coach counts more finely: per individual muscle, with exercises that only assist a muscle counted partially. Those are two different quantities, and the second one does not belong next to the figure of ten.
There is no fixed cycle. The app counts how many weeks in a row you have trained, and takes volume out as soon as your recovery check-in calls for it.
Try free for 14 days →