Upper lower split review: weekly sets and how to progress it

The usual mistake with an upper lower split is to stop counting at the days. Four sessions, every muscle trained twice a week, and the plan looks balanced on paper. Count the sets and it often is not. The app's own Upper / Lower template, run through the app's weekly volume rules, gives the back 30 weekly sets and the chest 9.5.

For progression, Double fits nearly every lift in it, with Top set + back-off for the two heaviest. The set counts and loads below are output from compute.mjs, the script that runs the app's own formulas.

What the template prescribes

The template is four sessions, 26 lifts and 85 working sets a week. The two upper days hold 23 working sets each, Lower A holds 20 and Lower B holds 19. On a weekday schedule the app places them on Monday, Tuesday, Thursday and Friday, so each half of the body gets at least two days off before it is trained again.

No lift appears twice in the week: the chest gets the bench press on Monday and the incline bench press on Thursday. Every session also runs from heavy to light, opening on the day's lowest rep range with three or four minutes of rest and ending on isolation work at 10 to 15 reps with a minute between sets.

Where the weekly sets land

The lower body and the back get plenty, and the chest does not. The count uses the app's rule for weekly sets per muscle: a working set credits each main muscle one set and each helper muscle half a set, and warm-ups credit nothing.

A meta-analysis of 15 studies found that more weekly sets meant more growth, with 10 or more sets per muscle as its top category, although that three-way comparison only reached a trend [1]. A later review of young, trained men suggested 12 to 20 weekly sets per muscle as a standard recommendation [2].

Against those two lines the quads, hamstrings and glutes land between 17.5 and 20.5. Biceps, triceps and front delts sit at 12 to 13, the bottom of the band. The chest at 9.5, the side delts at 9, the calves at 8 and the rear delts at 6.5 fall under both.

The back's 30 needs a caveat. Twenty of those sets come from lifts where the back is a main muscle. The other ten are half credits, because the app folds the lats, upper back and spinal erectors into one back muscle. A bent-over row lists two of them as main muscles and the third as a helper, so each set earns 1.5.

No session crowds one muscle either: the most any muscle receives in a day is 10.5 sets. A 2025 meta-regression, a preprint not yet peer reviewed, put the point where extra sets in a session stop reliably adding growth at about 11, counted with the same half credit for helper muscles [3]. Its authors ask for that figure to be read cautiously.

The cheapest fix for the chest is a swap. Replace Upper A's lat pulldown with three sets on the pec deck and the chest rises to 12.5 weekly sets, the back falls to 25.5 and the session stays at 23 sets. Three sets of lateral raises added to Upper B take the side delts from 9 to 12.

Twice a week is the reason to pick this split

An upper lower split exists to train each muscle twice a week inside four sessions. The evidence for twice over once is decent, and for anything finer it is thin. A 2016 meta-analysis of ten studies found a larger effect on muscle size at higher frequencies, 0.49 against 0.30, and concluded that training a muscle twice a week beats once [4].

The 2019 update, with 25 studies, found no significant difference between higher and lower frequencies once weekly volume was equated [5]. A 2024 meta-analysis of 14 studies and 392 participants found that split and full-body programs produced similar strength and muscle growth when volume was equated, and its authors told lifters to choose by preference [6].

The case for the split is practical. It suits a lifter with four dependable days who wants heavy barbell work and enough volume to grow in the same week, and a missed day still leaves the other session for that half of the body. The layout also matches how the app reads recovery, from hard sets in the last 48 hours and recent load in the last 72. By Thursday evening, Monday evening's upper session has left the first window and is at the end of the second. Our training split guide compares this layout with six others, and the push pull legs review counts a six-day plan the same way.

Which progression scheme fits each lift?

Double fits nearly every lift here, and Top set + back-off is worth choosing for the deadlift and the back squat if you rate effort. The app sets a scheme per lift from seven: Auto, Double, Top set + back-off, Load, Reps, RIR and Manual. Set to Auto, this template runs Double on all 26 lifts, because Auto picks Double whenever a rep range is at least two reps wide. Double holds the load until every working set reaches the top of the range, then adds the smallest step and the climb starts again.

The width of the range decides how much strength a lift must show before the load moves. Take an 80 kg lifter and the four opening lifts, each logged at RPE 8, a set stopped with two reps left. The table gives the estimated 1RM when the set lands at the bottom of its range and again at the top, at the same load. The estimates come from the model behind the one-rep max calculator.

The incline bench press has to gain 15.5 per cent in estimated 1RM before Double adds 2.5 kg, the app's default step for a barbell compound lift. The bench press needs 6.4 per cent. A wide range makes the first climb long, and later climbs are shorter. The lifter who earns the step to 87.5 kg on the bench press can expect about seven reps there, because the model's seven-rep load at RPE 8 for a 115 kg estimate is 87.6 kg.

Top set + back-off suits the two heaviest lifts. The template writes the deadlift as three sets of 3 to 5. Run it as one top set and two back-offs: 150 kg for three at RPE 8, then back-offs at 88 per cent of the top set, the app's default, which rounds to 132.5 kg. The model reads five reps at 132.5 kg as about RPE 6 for this lifter, four reps in reserve.

Keeping the heavy work on the opening lift has research behind it. In a meta-analysis of 21 studies where every set went to failure, heavier loads built more 1RM strength, while muscle growth was similar across loads [7]. For the lateral raise, where the next dumbbell up is a large share of the load, Reps fits better: hold the weight, add a rep at a time to a ceiling, then step up.

Where lifters stall on it

Progress is slower than on a plan that repeats its lifts, and Friday's deadlift is the weak point. Each lift appears once a week, so it gets one chance a week to progress. The app flags a stall after three sessions without progress by default, which on this template means three weeks.

The back is the one muscle with no clear day off inside the training week. It is credited 10.5 sets on Monday, 3.5 on Tuesday from the squat and the Romanian deadlift, and 10 on Thursday. Then Friday opens with the deadlift, a day after pull-ups and rows. If the deadlift stops moving while the bench press still climbs, look at Thursday first. Rating the top set helps: a top set logged as a grinder repeats the load the next week.

Upper days are long, and the arm work comes last. Skip the curls and the triceps work when a session runs over, and both muscles drop from 13 weekly sets to 7, all of it half credit from presses and pulls.

Logging it so the numbers stay honest

Log the ramp to each opening lift as warm-up sets. Warm-ups never count toward volume, so your weekly sets match the table. On the deadlift, mark the first working set as the top set and the next two as back-offs, and put an RPE on the top set. A set logged without RPE is treated as RPE 10, a true max, and the 150 kg triple would read as an estimated 162.7 kg where the rated set reads 173.8 kg.

The Bulgarian split squat is a per-side lift, logged left then right. The Top Sets app includes this template and counts weekly sets per muscle for 14 muscles as you log, so a chest stuck at 9.5 shows up in the first week.

Questions

Is an upper lower split good for building muscle?

Yes, when the weekly sets are there. Training a muscle twice a week beat once a week in a 2016 meta-analysis [4], and split and full-body programs produced similar muscle growth when volume was matched [6]. Count the sets each muscle gets before trusting any template.

Can I run an upper lower split three days a week?

You can, by rotating the four sessions through three training days. Weekly sets then swing. On the template reviewed here, a week with only one upper session gives the chest 4 or 5.5 sets, against 9.5 in a full week. If three days is your limit, a full-body plan spreads the work more evenly.

Is upper lower better than push pull legs?

Neither is better for growth when weekly sets are equal. A meta-analysis of 25 studies found no significant difference between training frequencies once volume was equated [5]. Upper lower trains each muscle twice in four sessions and push pull legs takes six, so choose by the days you can keep.