How Many Exercises Per Workout? (It's Set Math, Not a Number)

Four to six exercises covers almost every session, and across a full week most muscles need two or three genuinely different movements — not more. The count itself isn't a growth variable, though: exercise count is just the container you divide your weekly hard sets into, which means the question only has an answer once you've decided how many sets each exercise gets. Four exercises at three hard sets is 12 working sets. Eight exercises at one set each is eight. Same hour, same effort, and the first version builds more muscle. So the number worth arguing about is sets per muscle per week — exercise count is downstream of it.

How many exercises should you do per workout?

Four to six, on a session that trains two or three muscle groups. Multiply before you commit to it: exercises × sets = working sets, and working sets are what the growth research actually measures.

A push day with five exercises at three sets is 15 working sets — say six to chest, six across shoulders, three to triceps. Twice a week, chest lands at 12 — past the roughly 10 hard sets per muscle per week that Schoenfeld, Ogborn and Krieger's 2017 dose-response meta-analysis of 15 studies points at. Now the version people drift into: eight exercises because you wanted "more angles," one or two sets each because the clock ran out. Same 90 minutes, fewer hard sets per muscle. Adding an exercise without adding time converts sets into variety, and variety isn't the input.

Run two sets per exercise and you need more exercises to hold that weekly volume up — a real reason to sit at five or six instead of three. Cutting an exercise costs double what cutting a set does.

How many exercises per muscle group?

Two to three per muscle across the week, which usually means one or two in any single session.

Chest is the clean example: flat press and incline press, two or three sets each, twice a week. That's eight to twelve hard sets for chest out of two exercises. Back needs one vertical pull and one horizontal pull because they're different movements, not because two is a lucky number. Muscles your compounds already hit — triceps, front delts, biceps — usually need one dedicated exercise, not three.

Where people go over is the third and fourth movement for a muscle they care about. A decline press on top of flat and incline doesn't buy a new region of chest — it buys a third helping of a joint action you already did twice. Which is where the research gets interesting.

Does hitting a muscle from more angles actually build more muscle?

Sometimes yes, sometimes it costs you — and the difference is whether the variation is systematic or random. This is the part the "hit it from every angle" advice skips, and the literature on it is surprisingly clear.

The paper that settles the shape of the answer is Kassiano and colleagues' 2022 systematic review in the Journal of Strength and Conditioning Research — eight studies, combined sample of 241 young men. Their conclusion, almost word for word: some degree of systematic variation enhances regional hypertrophy and maximizes dynamic strength, while excessive, random variation may compromise gains. Then the line that should end most exercise-selection arguments: exercises that provide a redundant stimulus, and excessive rotation of exercises, "may actually hinder muscular adaptations." Not "add nothing." Hinder.

The evidence for the good half is Fonseca and colleagues' 2014 trial: 49 people, 12 weeks, twice a week, split across four training conditions — crossing varied-versus-constant exercise with varied-versus-constant load — plus a control group. Whole-quadriceps growth was similar everywhere (9.3% to 12.2% across both legs). The distribution wasn't. The two varied-exercise groups grew all four heads of the quadriceps; the constant-exercise groups missed the vastus medialis or the rectus femoris. And the group that varied exercise while holding load constant produced the largest strength gains of the four. The title says it plainly: changes in exercises are more effective than changes in loading schemes to improve muscle strength.

The evidence for the bad half is Baz-Valle and colleagues' 2019 study in PLoS One, with Brad Schoenfeld on it. Twenty-one trained men, eight weeks, 3 sets of 6 exercises, four sessions a week — one group on a fixed exercise selection, the other with its exercises and rep ranges randomized every session by an app. Bench and squat 1RM rose sharply in both, with no difference between groups, and vastus lateralis and rectus femoris thickness came out the same way. On the vastus intermedius there was no between-group difference either, but only the fixed group grew significantly from baseline. The randomized group's one clear win was psychological: a significant, moderate rise in intrinsic motivation, while the fixed group's drifted down.

Read the three together and the verdict is a split, not a debunk. Choosing two or three exercises that load a muscle differently on purpose is worth doing. Rotating exercises because you're bored is an adherence tool — a real one, since a program you enjoy is a program you keep running — but it isn't a growth multiplier, and randomizing session to session is the version the data likes least.

When is an exercise redundant?

The test, in one sentence: if you can't name the joint action, the muscle length, or the region of the muscle that the new exercise trains and your existing ones don't, it's redundant.

• Flat bench, then incline bench: not redundant. The shoulder angle changes, which changes which part of the chest takes the load. Two real positions.
• Then a decline press or a third machine press: redundant. Same joint action, third time, no new position. Those sets would do more as extra sets on the two presses you already picked.
• Squats, then a hip hinge: not redundant. A squat is knee-extension dominant and barely trains the hamstrings at the hip. A Romanian deadlift or leg curl is a different job entirely.
• Pull-ups, then lat pulldowns: redundant. Same movement, different implement. A row isn't: vertical and horizontal pulling put the shoulder blades in different places.
• Dumbbell curl, then cable curl: redundant. Incline curl and preacher curl aren't, because one trains the biceps in a stretched position and the other in a shortened one.

Kassiano's authors frame the selection rule as "applied anatomical and biomechanical constructs," which is academic for: pick your exercises by what they mechanically do, not by how many of them there are. Two exercises chosen that way beat four chosen off a list.

Is 7 exercises per workout too much? Is 10?

Neither number is too much on its own — time-and-fatigue arithmetic answers both, so run it instead of asking for a rule.

Seven exercises at two hard sets each is 14 working sets — at two to three minutes of rest, a normal 60-minute session and a defensible one. Seven at four sets each is 28 working sets, close to two hours, and the back half gets performed by a tired person. The exercise count didn't change. The answer did.

Ten is where it usually breaks: 30 working sets in one session, realistically 100+ minutes done properly. The two ways people make that fit are cutting rest and cutting effort, and both tax the sets they already had. If you want ten movements, run two sessions of five — same exercises, twice the recovery, and how close to failure each set ends stays where it has to be for the set to count.

Fatigue also lands unevenly: whatever sits last on the list gets your worst effort every week, so a muscle that matters goes early. Bigger lever than an eighth movement.

How do I pick the exercises for a session?

By what the session already covers, then by what's missing — and the log decides, not the feeling.

I'm 37, I work full-time as a software engineer, and I train four days a week: Push, Pull, Legs, Upper. The default I teach inside that split is deliberately boring — 3 sets of 8-12 reps at 1-2 reps in reserve — and my position on sets has been on the record all along: two or three genuinely hard sets beat four junk sets. Most guys don't have a volume problem, they have an effort problem. I'd rather see five exercises done hard than eight done half-assed.

What I won't do is hand you my personal exercise list as if it were the finding. My selection isn't the transferable part; the selection rule is. Pick the movements first, get the count from the sets, and stop adding when you can't name what a new exercise does that the others don't.

The measurement is the part almost nobody runs. You don't catch a redundant exercise by feel — it feels great, that's the problem. You catch it in the logbook: if the weight on one exercise hasn't moved in a month while everything around it climbs, it's either your priority or a passenger. I once ate about 300 surplus calories a day because the apples I bought weighed more than the app's "standard apple." I only caught it because I weigh things. Same principle here: write it down, watch the trend, change one thing.

If you want the split, the exercise list, and the set-and-rep spec as one system that survives a real work week, that's The Momentum Method.

And when you do add an exercise, add it for a reason you can say out loud. "More angles" isn't one. "My upper chest is behind and nothing I run puts the shoulder in flexion" is.