Training frequency: how to split weekly sets across sessions

Sixteen weekly chest sets can be logged as one session of 16, two of 8, or three of 5 to 6. For muscle size, the research finds little to separate them once the weekly total is equal [1]. That makes training frequency, the number of times a week a muscle is trained, mostly a way to keep each session small enough that its last sets are still worth doing. A 2025 analysis of per-session volume gives a working rule: split a muscle's weekly sets so that no session holds more than about 11 of them [2]. For 16 sets, that is twice a week.

The set credits and loads below come from compute.mjs, the script that runs the app's own formulas. Every other figure carries a citation to the paper it came from.

Frequency is a way to deliver volume

How often a muscle is trained matters much less for size than how many hard sets it gets in a week. Our training split guide walks through the two meta-analyses most lifters have heard of, so this is the short version. In 2016, ten studies pointed to twice a week beating once, and the authors could not say whether three times beat twice [3]. By 2019 there were 25 studies. With weekly volume equated, higher and lower frequencies no longer differed significantly, and the authors' advice was to choose a frequency per muscle by preference [1].

The largest analysis so far splits the answer by goal. Across 67 studies and 2,058 participants, frequency's effect on muscle size was compatible with negligible, while strength rose with frequency, with diminishing returns [4]. An older meta-analysis of 22 strength studies saw the same rise, from an effect size of 0.74 at one session a week to 1.08 at four or more. The rise disappeared in the studies that equated volume, and most of the studies it pooled used untrained participants [5].

Read together, extra sessions help mainly by carrying extra sets and, for strength, extra practice. Any benefit from spreading the same sets over more days is small enough that the bigger analyses cannot pin it down.

Has anyone tested one session of 16 against two of 8?

Yes, once, and the study was small. Brigatto and colleagues assigned 20 trained men to train each muscle group once a week with 16 sets, or twice a week with 8 sets a session, for eight weeks [6]. Both groups improved. No significant difference between them showed up in squat or bench press strength, in muscle thickness at the triceps, elbow flexors or quads, or in muscle endurance. Some effect sizes leaned towards the twice-a-week group, which the authors read as a possible slight benefit of the higher frequency.

The same laboratory then pushed frequency much further. Eighteen well-trained men followed either a split that trained each muscle group once a week or a whole-body plan that trained each one five days a week, again for eight weeks [7]. Strength did not differ between the groups. Muscle thickness rose more in the five-day group at two of the three sites measured, one of them the vastus lateralis of the quads.

Nine or ten lifters a group over eight weeks cannot settle a question this fine. The first trial does show that 16 sets in one session still worked for trained men, and in neither trial did the spread-out group come out behind.

Where extra sets in one session start to count for less

At about 11 sets for one muscle, on the best current estimate, and the estimate is rough. A 2025 meta-regression from the group behind the 67-study analysis looked at sets per session instead of sets per week [2]. More sets in a session meant more growth, with diminishing returns. The authors then looked for what they call the point of undetectable outcome superiority: the per-session volume past which one more set no longer had a better-than-even chance of adding a difference large enough to detect. For muscle size it sat at about 11 sets. For strength it sat at about 2 sets of the lift being tested.

Two cautions come with those numbers. The paper is a preprint and has not been peer reviewed. Studies with very high per-session volumes are also scarce, so the authors cannot say whether growth keeps creeping up past that point or starts to fall, and they ask for the figure to be interpreted cautiously [2]. In their model, sets past the eleventh still counted. Each one counted for less, with more uncertainty about how much.

The unit matters too. The 11 is in fractional sets: a set of a lift that trains the muscle directly counts as one, and a set where the muscle only assists counts as half [2]. The app gives a secondary muscle the same half credit when it counts weekly sets per muscle for 14 muscles, and how many sets per week per muscle explains where it comes from. A session's set credits in the app are already in the study's unit.

The app does not discount late sets. The sixteenth chest set of a session earns the same full credit as the first, so the weekly count reads 16 however the sets arrived. To see how they arrived, count the credits one session at a time.

Sixteen chest sets, three ways

Take an 80 kg lifter whose week holds 6 sets of barbell bench press, 6 of dumbbell incline press and 4 on the pec deck. In the app's exercise library the bench press lists the chest as primary, with the triceps and front delts as secondary. The incline press lists the upper chest and front delts as primary and the triceps as secondary, and the upper chest is counted as chest. The pec deck lists the chest alone. Under the app's volume rules that week credits the chest 16 sets, the front delts 9 and the triceps 6.

Split in two, each day gets 3 sets of each press and 2 on the pec deck. Split in three, each day gets 2 sets of each press, with 2 pec deck sets on the first day and 1 on the others.

Only the single session crosses the 11-set point, and it crosses by 5. With the lifts done in that order, those 5 are the last incline set and all four on the pec deck, done on a chest that already has 11 sets in it. The two-session layout keeps both days 3 sets under. The three-session layout has more room than 16 sets need, which starts to matter only when the weekly total grows.

Frequency also changes what goes on the bar. Say the lifter benches 85 kg for 8 at RPE 8, the effort rating where 10 means no reps left and 8 means about two in reserve. The app's model reads that set as an estimated 1RM of 115 kg, the figure the one-rep max calculator gives for the same set. One bench day a week means one rep target. Three bench days leave room for three: the model puts five reps at RPE 8 at 93.3 kg and twelve at 74.6 kg, which round to 92.5 kg and 75 kg on a bar with 2.5 kg steps.

So Monday is two sets of five at 92.5 kg, Wednesday two of eight at 85 kg and Friday two of twelve at 75 kg. Two bench sets a day is also where the per-session analysis placed the point of diminishing returns for strength [2].

The cost is warm-ups. The app's ramp to 85 kg is four sets: the bar for 8, 42.5 kg for 5, 60 kg for 3 and 72.5 kg for 2. One bench day pays that once. The three-day version needs four sets to reach 92.5 kg, four more to reach 85 kg and three to reach 75 kg, which makes 11 warm-up sets a week against 4. Warm-ups never count toward volume, so none of them touch the 16. They still take time.

What the recovery view shows for each layout

The app reads recovery from two rolling windows: hard sets in the last 48 hours and recent load in the last 72 hours. Both windows count set credits, so the three layouts leave different traces even though the calendar-week count shows 16 chest sets for each. The muscle recovery time guide covers the windows in full. Here is how they fill when every session is at 6 pm.

A muscle's readiness label is worked out from the sets inside those windows, and sets inside 48 hours weigh on it more. The single session puts all 16 chest sets into both windows at once. They leave the 72 hour window at 6 pm on Thursday, and the chest then reads "No recent load" for four days. With two sessions the windows never overlap: Monday's 8 sets drop out on Thursday evening just as the next 8 go in.

The three-session layout is the only one where sessions stack. On Thursday morning the 48 hour window holds Wednesday's 5 chest sets, and the 72 hour window holds 11: those 5 plus Monday's 6. The chest never returns to "No recent load" between sessions, and it never has more than 6 sets in the 48 hour window.

None of these pictures is a verdict on growth. The windows count sets in time, and the research above gives most of the credit to the weekly total. The view earns its place when you fit other work around the pressing. The front delts and triceps ride along with every chest session: by Thursday morning in the three-session layout, the 72 hour window holds 6 front delt sets and 4 triceps sets from chest work alone.

An overhead press day on Thursday lands on shoulders already carrying that load, and it adds to the chest as well. Three sets of overhead press credit the front delts and side delts 3 sets each and the triceps 1.5. The lift lists the upper chest as secondary, so the chest gets 1.5 too: a fourth day with chest credit in a week planned around three.

How to choose a frequency for a muscle

Start from the weekly number, then divide. Decide how many weekly sets the muscle gets, divide by 11 and round up. The result is the fewest sessions that keep each one at or under the point where extra sets start to count for less [2].

The table is arithmetic on one estimate from one preprint, so treat it as a starting layout. Count in set credits, with half sets for assisting lifts, or the sessions will look smaller than they are.

If one session would do, two is still the safer default. Ten weekly sets fit inside a single day, yet the 2016 meta-analysis found twice a week ahead of once [3], and the effect sizes in the 16-against-8 trial leaned the same way [6]. Space the days 48 to 72 hours apart, so the first session has left one or both of the app's recovery windows before the next begins.

For strength, lean the other way: more days, fewer sets a day. Strength rose with frequency in the largest analysis [4], and its returns per session thinned after about 2 sets of the lift [2]. The authors of the per-session paper suggest low per-session volume on roughly two to three days a week for strength over the short to medium term [2]. The three-day bench layout above fits that description.

Then check the layout against your week. A missed session costs all 16 chest sets in the first layout, 8 in the second and 5 or 6 in the third. Pick the frequency you can repeat for a whole block. Top Sets counts weekly sets per muscle for 14 muscles as you log, so you can see whether the week delivered what the layout promised. The weekly sets per muscle calculator runs the same count on a planned week.

Mistakes when spreading sets

Counting gym days as frequency is the first. The findings above are about how often a muscle is trained, which the 2016 review separates from how many sessions a lifter does [3]. Six days of push, pull and legs is a frequency of two for the chest. Five days of one body part a day is a frequency of one.

Adding a day and its sets together, then crediting the day, is the second. A lifter who goes from two chest days of 8 sets to three has added 8 weekly sets. Where studies did not equate volume, higher frequencies did come out ahead, modestly [1], and the strength analysis traced its frequency effect mainly to volume [5]. Change the layout or the total, one at a time.

Leaving half credits out of the session count is the third. The single 16-set chest day also puts 9 sets on the front delts and 6 on the triceps before any direct shoulder or arm work. Follow it with 8 sets of cable pushdowns and the triceps have had a 14-set session, 3 over the point where returns thin.

Treating 11 as a wall is the fourth. A session of 12 or 13 sets that fits the week is not worth breaking up on the strength of one preprint. The estimate marks where returns thin, and its authors say plainly that the data above it are scarce [2].

Questions

Is training a muscle once a week enough to build muscle?

Yes, if the weekly sets are there. With volume equated, a meta-analysis of 25 studies found no significant difference in growth between higher and lower frequencies, and a trial of 16 sets once a week against 8 sets twice found similar results in trained men. The risk is the long session: past about 11 sets for one muscle, each extra set appears to add less.

Is training a muscle three times a week better than twice?

For size, not measurably. A 2016 meta-analysis could not tell three from two, and later analyses found frequency's effect on growth compatible with negligible once weekly sets were matched. Three sessions earn their place when the weekly total is high enough that two sessions would each pass about 11 sets, or when strength on one lift is the goal.

How many sets per muscle in one session is too many?

There is no proven ceiling. One preprint meta-regression placed the point of diminishing returns for muscle size at about 11 sets per session, counting assisting lifts as half a set. Sets beyond that still added growth in the model, only less, and data on very large sessions are scarce. If a session runs past 11 for one muscle, move the extra sets to another day.