How to pick a training split for 3, 4, 5 or 6 days a week

Three days a week of the same 17-set full-body session gives the chest 9 weekly sets and the back 16.5. That one line holds the whole method for choosing a training split: start from the number of days you can train every week, pick a split that hits each muscle at least twice, then count the sets it actually delivers per muscle. Frequency matters less than most split debates suggest. Weekly sets matter more, and a split is mostly a way of fitting them into the days you have.

Every weekly set count below comes from compute.mjs, the script that runs the app's own volume rules: a working set credits its main muscles one set and its helper muscles half a set, and warm-ups count for nothing. The exercise-to-muscle mapping is the app's stock catalogue. Change the exercises and the numbers change, which is the point of counting.

Days per week decide the split

Your available days narrow the choice before anything else does. A split that needs six sessions fails the first week you can only make four, and a missed day in a body-part split means a muscle goes untrained for a fortnight.

This is how the common splits line up against a week:

The frequency column is read straight off the templates: a muscle counts as trained on a day when one of that day's exercises lists it as a main muscle. Helper credit, like the triceps work in a bench press, adds more touches on top.

Pick the row for the number of days you can keep for a whole training block, not the number you managed in your best month. A block is the run of weeks you plan to repeat before you change the program. If your week is unpredictable, take the row below what you hope for. Full body and upper/lower both survive a missed session better than a bro split, because every muscle has another day coming.

How often should a muscle be trained?

Twice a week beats once a week when the volume is the same, and past twice the difference shrinks to nothing measurable. A 2016 meta-analysis of ten studies found that training each muscle at least twice a week produced larger growth than once, with effect sizes of 0.49 against 0.30 [1]. The same authors could not say whether three times beats twice.

A 2019 update with 25 studies went further. When weekly volume was equated, higher and lower frequencies produced no significant difference in growth, including in trained lifters and when upper and lower body were analysed separately [2]. The authors concluded that for a given weekly volume, frequency per muscle can be chosen by preference. Higher frequencies came out ahead only in studies where they also added volume.

Strength follows the same pattern. Across 22 studies, effect sizes rose from 0.74 at one session a week to 1.08 at four or more, but the advantage disappeared once volume was equated [3]. Frequency is a delivery method for volume.

Weekly volume itself does matter. In a 2017 meta-analysis, each extra weekly set per muscle was associated with a small additional gain in muscle size, and the group doing 10 or more weekly sets per muscle did best, although that three-way comparison only reached a trend [4]. The site's guide to weekly sets per muscle covers where the targets come from and how to adjust them.

There is one direct comparison in trained men worth knowing. Over ten weeks, a full-body program and a split program both improved every strength and muscle-thickness measure. The split group gained more thigh muscle thickness and the full-body group did better on one bench press strength test [5]. Twenty-one lifters over ten weeks is a small study, so read it as "both work", not as a winner.

Seven templates, counted

The templates below use three working sets per exercise unless stated, built from common lifts. The table shows the weekly sets the app's rules credit to five muscles.

What each template contains, so you can check the counts against your own:

• Full body: squat, bench press and pull-up for 3 sets each, then Romanian deadlift, lateral raise, curl and pushdown for 2 each, the same session three times.
• Upper/lower: bench press, cable row, overhead press, pull-up, lateral raise, curl and pushdown on upper days. Squat, Romanian deadlift, leg press, leg curl and calf raise on lower days.
• Torso/limbs: bench press, cable row, overhead press, pull-up, pec deck and lateral raise on torso days. Squat, Romanian deadlift, leg extension, leg curl, curl, pushdown and calf raise on limb days.
• Push/pull/legs: a push day of bench press, incline bench, overhead press, lateral raise and pushdown, a pull day of pull-up, cable row, face pull and curl, and the lower-day session from upper/lower as the leg day.
• Arnold split: chest and back together (bench, incline bench, pull-up, cable row), shoulders and arms together (overhead press, lateral raise, face pull, curl, pushdown), then legs.
• Bro split: 4 sets per exercise, with chest, back, shoulders, legs and arms each getting their own day.

The 3-day options

Full body three times a week is the strongest 3-day choice on these numbers. Every main muscle is trained three times, the back reaches 16.5 sets because the pull-up, the squat and the Romanian deadlift all credit it, and the whole thing costs 17 working sets a session. Its weak spots are the chest and quads at 9, just under the 10-set mark, and the calves and rear delts at zero because nothing in the session trains them. Adding one set of bench and one of squat to each session would close the first gap.

Push/pull/legs run once a week is the weakest option on the list. Each muscle is trained once, and the quads get 6 weekly sets. It is a fine template for a lifter whose real goal is three short sessions, but it is a poor way to grow.

The 4-day options

Upper/lower and torso/limbs both train every main muscle twice and give the quads 12 weekly sets. The difference is where the arms go. Torso/limbs moves curls and pushdowns to the leg days, which frees room on torso days for a pec deck, and the chest climbs from 9 to 15 sets. Arms are trained four times a week in that layout, twice directly and twice as helpers.

Upper/lower leaves the chest at 9 because the upper day is already 21 working sets long. If chest is a priority on upper/lower, swap the cable row for a second press on one of the two upper days, or add two sets of pec deck and accept a longer session.

The 5-day bro split

The bro split puts 14 sets on chest and triceps and only 8 on the quads and side delts, and most muscles are trained once a week. On the frequency evidence above, it loses nothing when volume is matched [2]. The practical problem is the long single session: 20 working sets on leg day, with the last few done on tired legs. It suits a lifter who likes long, focused sessions and never misses a day.

The 6-day options

Push/pull/legs twice and the Arnold split use the same exercises, so the weekly counts are identical: 15 for chest, 21 for back, 12 for quads. What differs is the shape of the week. The Arnold split trains the chest on four days, two direct and two as half-set helper credit from the overhead press, while push/pull/legs trains it on two. With volume equal, the research gives no reason to prefer one over the other [2]. The push/pull/legs version is reviewed in detail in our push pull legs program review.

A worked example on upper/lower

Take a 90 kg lifter on the 4-day upper/lower template. One upper session of 21 working sets credits the chest 4.5 sets: 3 from the bench press and 1.5 from the overhead press, which lists the upper chest as a helper. The back gets 6, and the biceps and triceps get 6 each. Two upper days make the weekly 9 chest sets in the table.

The lifter benches 80 kg for 6 on Monday at RPE 8, which means they finished the set with about two reps left. The app reads that set as an estimated 1RM of 101.8 kg, using the same model as the one-rep max calculator.

The second upper day is a chance to train the same muscle at different reps rather than repeat Monday. The model says ten reps at RPE 8 from a 101.8 kg max is 69.2 kg, which rounds to 70 kg on a barbell with a 2.5 kg step. So Thursday is three sets of ten at 70 kg, and both days feed the same weekly chest count.

Logging the two sessions under one program is what makes the count possible. The app shows weekly sets per muscle for 14 muscles, so a lifter can see the 9 chest sets as they land rather than adding them up by hand. If you are moving a split over from another logger, the guide to switching from Strong or Hevy covers importing your history.

Mistakes that break a split

Counting sessions instead of sets is the most common one. A 6-day program sounds like more training than a 4-day one, yet push/pull/legs run once over three days gives the quads half what upper/lower does over four. Count weekly sets per muscle, not days in the gym.

Ignoring helper credit comes next. A lifter who writes 15 sets of direct triceps work on top of a pressing-heavy split can end up far higher than intended, because every bench and overhead press set adds half a set. The bro split's 14 triceps sets come from 8 direct sets plus the pressing on chest and shoulder days.

Then there is picking the split for the week you wish you had. A missed leg day in a bro split is a week with no quad training at all. A missed lower day in upper/lower still leaves the other one.

Last comes changing split every few weeks. Frequency differences are small once volume is matched, so the switch rarely buys anything, and it resets the comparison between sessions that progression depends on. Run one split for a full block, then judge it by whether your logged loads and estimated 1RMs moved.

Questions

Is full body or a split better for hypertrophy?

Neither wins when weekly volume is the same. A 2019 meta-analysis of 25 studies found no significant difference in growth between training a muscle more or less often once volume was equated [2]. Pick the layout that gets each muscle 10 or more hard weekly sets in the days you can train, and that you can repeat without dreading a session.

How many days a week should I train for muscle growth?

As many as you can keep every week for a whole block. Three full-body sessions can train every main muscle three times, and four upper/lower sessions train each one twice with more room per session. Frequency on its own matters little once volume is matched, so the deciding factor is the number of weekly sets you can fit and recover from.

Is a bro split bad?

No, but it is fragile. Training each muscle once a week with enough sets grew muscle about as well as higher frequencies in volume-equated studies [2]. The cost is long single sessions, with fatigue building through the later sets, and a missed day that leaves a muscle untrained for two weeks. Earlier work found twice a week ahead of once [1].

Do warm-up sets count toward weekly sets in a split?

No. The app counts working sets only: a warm-up set credits nothing, a working set credits its main muscles one set, and a helper muscle gets half a set. That is why the templates above count only working sets, and why a ramp of light singles before your top set never inflates a muscle's weekly total.