How Many Hard Sets Per Muscle Per Week — and Why Your Count Lies

Most lifters should do roughly 10-20 hard sets per muscle group per week, where "hard" means a working set taken within about 0-3 reps of failure. That range is the practical consensus from the volume research. But there is a catch the guidelines never mention: every one of those studies counted hard sets, and you are almost certainly counting attempted sets. If half of yours stopped short, your real weekly volume is half of what your logbook says, and your muscles are getting beginner-level stimulus on an advanced-level schedule.

I see this every week. A lifter shows me a program with "16 sets for chest," frustrated that nothing is moving. We watch the sets. 4 of them end with the bar still flying up and 4 clean reps left in the tank. That is not 16 hard sets. That is 8, padded with junk. The number on the page is real. The stimulus it implies is fiction.

Here is how the count actually works, where it lies, and how to find out which of your sets counted, including a 1-week worksheet for auditing your own numbers.

How many hard sets per muscle per week should you do?

For most trained lifters, 10-20 hard sets per muscle per week is the productive window, with bigger muscles tolerating the high end and smaller ones topping out lower. This is not a precise dose. It is a range, and the research behind it is softer than the confidence with which it gets repeated.

The anchor is Schoenfeld, Ogborn and Krieger's 2017 dose-response meta-analysis. Pooling 15 studies, they found a graded relationship: each additional weekly set per muscle raised the hypertrophy effect size by about 0.023, which they translated to roughly 0.37% extra muscle per set, and 10+ weekly sets outperformed 5-9, which beat fewer than 5. More volume, more growth, up to a point, with diminishing returns per set. Note the magnitude. That is a gentle slope, not a linear ramp you can ride to 40 sets.

The other half of the story is frequency, and it matters for how you hit your number. A separate Schoenfeld meta-analysis found that, when weekly volume is equated, training a muscle twice per week produces more growth than cramming the same sets into 1 session (effect size 0.49 vs 0.30). So 14 sets for back is not best done as 1 brutal 14-set marathon; it is better split across 2 sessions of 7. More on that below.

What is a "hard set," and why does it change the number?

A hard set is a working set taken close enough to failure, within roughly 0-3 reps, that it actually stimulates growth; anything easier barely counts toward your weekly total. This single definition is what separates "16 sets logged" from "16 sets that did something."

Why the proximity requirement? Because not every rep in a set grows muscle. The growth-driving reps are the last handful, where 2 things finally coincide: nearly all your motor units are recruited and the bar is moving slowly enough to put high mechanical tension on each fiber. Early reps of a submaximal set are fast and easy; they accumulate fatigue without much stimulus. If you stop a set 5 reps shy of failure, you never reach the part of the set that matters. You did the costly reps and skipped the productive ones. I dig into which reps actually pay off in effective reps explained.

The proximity-to-failure literature confirms you do not have to grind all the way to failure to collect those reps. Refalo and colleagues' 2022 meta-analysis found no significant hypertrophy advantage to training to momentary failure over stopping short when volume was equated (effect size 0.12, with a confidence interval that crossed zero, −0.13 to 0.37). But the same review's velocity-loss analysis shows you do have to get close: the effect size climbed from 0.20 at low velocity loss (under 20%) to 0.39 at moderate loss and 0.42 at high loss. The jump from "easy" to "moderately hard" nearly doubled the effect. The grind from "moderately hard" to "absolute failure" barely moved it. So a hard set is one taken to about 1-3 reps from failure: not an all-out grind, but unmistakably not easy. How close to failure should you train breaks that target down by goal.

Counted sets vs effective sets: where your count lies

Your logbook counts every set you attempted. The research counts only the sets that were genuinely hard. Those 2 numbers are rarely the same, and the gap is where most stalled progress hides.

This is the core deception. "Volume" as most people track it is sets × reps × weight, or just "sets logged." That arithmetic cannot tell a near-failure set from one stopped 4 reps early. They look identical on paper: same exercise, same load, same rep count. But one delivered a full dose of stimulating reps and the other delivered almost none. When a study prescribes "12 sets per muscle," researchers are supervising effort; sets go to a controlled RIR, often with a coach calling it. Your self-directed 12 sets carry no such guarantee.

The fix is not to log harder-sounding words. It is to make the count honest. A leaner program of genuinely hard sets out-grows a sprawling one full of easy ones, and costs less recovery. If you suspect you are padding, junk volume walks through how to spot the wasted sets.

How does under-effort silently halve your real volume?

It halves your volume because the cost of a set lands on your logbook the instant you rack the weight, but the benefit only accrues in the final reps you may never reach, and you are a bad judge of whether you reached them. The result: the count goes up, the stimulus does not.

2 well-documented facts make this worse than it sounds. First, lifters systematically misjudge how close to failure they are. In a study of resistance-trained people, participants who believed they had hit failure, 0 reps left, could actually complete about 2 more reps on average (and in one of the experiments, nearly 3). The pattern holds across the proximity-to-failure literature: even trained lifters under-predict their reps-to-failure, and accuracy only improves when they are already close to failure, using heavier loads, or in later sets. So the lifter who calls a set "2 RIR" is often really at 4 or more. They think they did a hard set. They did half of one. Why your reps-in-reserve estimate is wrong goes through that error in detail.

Second, that error compounds across a session. If 4 of your 8 chest sets each stopped 4 reps early, you did not do "8 sets minus a little." You did 4 fully effective sets plus 4 near-empty ones: functionally 4-6 effective sets out of 8 logged. Scale that across a week and the "16-set" chest program is delivering the stimulus of a beginner's 8. You are then mystified about why you are not growing, which is the exact gap I unpack in why am I not building muscle. The number lied. It always quietly does, because effort is invisible and the logbook only records that a set happened.

Should you split volume across more sessions?

Yes. For most muscles, splitting your weekly hard sets across 2 (or more) sessions beats doing them all at once, because volume-equated research favors higher frequency and because set quality decays late in a long session. Frequency is not magic on its own; it is a tool for fitting in more good sets.

When Schoenfeld's frequency meta-analysis equated weekly volume, twice-a-week training beat once-a-week (effect size 0.49 vs 0.30). The likely mechanism is partly practical: the eighth straight set for a muscle is rarely as hard as the second. Fatigue erodes the load you can move and the reps you can grind, so a 14-set marathon produces a tail of degraded, semi-junk sets. Split it into 2 × 7 and every set starts fresher, so more of them actually clear the "hard" bar. Frequency's real benefit may be that it protects set quality, which loops straight back to the effective-set problem above.

Practically: if a muscle needs 10-20 hard sets, run it twice a week at 5-10 sets per session. Going 3 times can work for large, recoverable muscles, but 2 is the reliable default that most evidence supports.

How do you measure which sets actually counted?

You measure it by checking effort objectively instead of trusting feel, and the most measurable proxy for proximity to failure is how much your reps slow down. As a set approaches failure, bar speed drops in a predictable way, and that decay is something you can actually quantify.

The velocity signal is real and reasonably tight. Sánchez-Medina, González-Badillo and colleagues found a strong relationship between velocity loss and the percentage of a set's reps completed (R² around 0.93-0.97 across bench press and squat), and a controlled bench-press study showed lifting velocity decreased far more after sets to failure (−25%) than at 1 RIR (−13%) or 3 RIR (−8%). In other words, rep slowdown tracks how deep you actually went, which is precisely the thing your subjective sense gets wrong. Why your reps slow down at the end of a set explains the mechanism, and velocity loss thresholds explained covers what cutoff to stop at.

One caveat before you treat velocity as gospel: the relationship between slowdown and reps in reserve shifts by exercise, load and set number. It is an objective second opinion on feel, not a replacement for it. That is still a large upgrade, because feel is the thing that has been lying to you.

Here is the quick version, for your single most-stalled muscle this week:

1. Pick the muscle. Take its weekly set count off the logbook; call it your claimed volume.
2. Film 3 sets, side-on. Watch the last 2 reps. If the bar is still moving fast and smooth, you had reps to spare; that set was not hard.
3. Recount honestly. Tally only the sets where the final rep clearly slowed and you would have ground to failure within 1-2 more. That is your effective volume.
4. Compare. If effective is far below claimed, your problem was never the program. It is effort.
5. Fix intensity before adding sets. Push your existing sets 2 reps deeper before you add a single new one. It is the cheapest upgrade in lifting.
6. Re-check after 2 weeks. Effort drifts back toward comfortable. Audit again.

The full version, for every muscle at once, is the worksheet below.

The real volume audit: a 1-week worksheet

Copy this into your notes or print it. For 1 week, fill in a row for every muscle you train. Warm-ups do not go in any column.

How to fill each column:

• Sets counted. Every working set you logged for that muscle this week. This is your claimed volume, the number you have been comparing against 10-20.
• Sets that ended with a visible slowdown. Sets where the last rep was clearly slower than the first few. Film it side-on, or have a training partner call it. A slowdown you felt does not count; it has to be visible, because feel is the thing that lies.
• Real hard sets. The slowed-down sets you would have failed within 1-3 more reps. A single slightly slower rep followed by an easy lockout stays in column 3 and does not make it into column 4. This is your effective volume, and it is the only column the research is talking about.

A worked example, using the chest program from the opening: 16 sets counted, 10 that showed some slowdown, 8 that were genuinely 1-3 reps from failure. Effective volume 8, exactly half of what the logbook claimed. That "high-volume" chest plan was a low-volume plan with a long warm-up.

How to read your results:

• Column 4 is under 10 while column 2 says 12-20. The program was never the problem. Make the existing sets harder before you add one: push each set 2 reps deeper for 2 weeks, then re-run the audit.
• Column 4 is already 10-20, every set genuinely hard, and progress is still flat. You have earned the right to add load rather than sets; when to add weight covers the trigger. If instead everything feels heavy, sleep is poor and the slowdown now shows up on your first sets, the answer is fewer sets for a week, not more. Do you need a deload week has a checklist for telling the 2 apart.
• Column 4 is well over 20. That is more hard sets than the research supports for most muscles. Trim the count and keep the effort; the dose-response curve has already flattened up there.
• The gap between column 2 and column 4 is the number to fix. It is your effort error, and it will come back. Effort drifts toward comfortable, so re-run the audit every few weeks.

Where Riven fits

Riven is the Apple Watch app that scores muscle failure. It reads your wrist motion and stays quiet during the set; about three seconds after you rack the weight you get the exercise, the rep count and a 0-100 score. Reps slow down as a muscle nears failure, and that is what the score is built on. No barbell clip, no camera, no extra hardware. It shows you which set was the real one.

FAQ

How many sets per muscle per week is best for muscle growth?

For most trained lifters, 10-20 hard sets per muscle per week, with bigger muscles tolerating the upper end and smaller ones the lower. "Hard" means within about 0-3 reps of failure. Beginners grow on fewer (6-10), and the dose-response curve flattens with diminishing returns per added set, so more is not endlessly better.

Do all my logged sets count toward that 10-20?

No. Only sets taken close to failure count meaningfully. Volume research supervises effort, so its set counts are all hard sets. If your sets stop several reps early, they barely contribute, and your real weekly volume can be half your logged number even though the logbook looks full.

Do warm-up sets count toward weekly volume?

No. A warm-up is, by design, stopped far from failure, so it never reaches the slow, high-tension reps that drive growth. Count only working sets taken within about 0-3 reps of failure. Leave warm-ups out of every column of the audit.

Is it better to do all my sets in one session or split them across the week?

Split them. When weekly volume is equated, training a muscle twice per week beats once (effect size 0.49 vs 0.30), partly because set quality decays late in a long session. Run 10-20 weekly sets as 2 sessions of 5-10 rather than 1 marathon.

How do I know if a set was actually hard?

Watch the last 2 reps: if the weight is still moving fast and smooth, you left reps in the tank. Rep slowdown is the most measurable failure cue; lifting velocity drops about 25% by failure versus 13% at 1 RIR. Riven is the Apple Watch app that scores muscle failure; it scores each set the moment you rack it.

Why am I not growing if I'm doing 15+ sets a week?

Most likely because those sets are not as hard as you think. Trained lifters under-predict reps-to-failure by about 2 on average, so a "15 hard sets" week is often 7-8 effective sets plus junk. Fix effort before adding volume: push existing sets 2 reps deeper.

Sources

• Schoenfeld, Ogborn & Krieger (2017), Dose-response relationship between weekly resistance training volume and increases in muscle mass: A systematic review and meta-analysis, Journal of Sports Sciences — https://pubmed.ncbi.nlm.nih.gov/27433992/
• Schoenfeld, Ogborn & Krieger (2016), Effects of Resistance Training Frequency on Measures of Muscle Hypertrophy: A Systematic Review and Meta-Analysis, Sports Medicine — https://pubmed.ncbi.nlm.nih.gov/27102172/
• Refalo, Helms, Trexler, Hamilton & Fyfe (2022), Influence of Resistance Training Proximity-to-Failure on Skeletal Muscle Hypertrophy: A Systematic Review with Meta-analysis, Sports Medicine — https://pmc.ncbi.nlm.nih.gov/articles/PMC9935748/
• Rodríguez-Rosell, Sánchez-Medina, González-Badillo et al. (2020), Relationship Between Velocity Loss and Repetitions in Reserve in the Bench Press and Back Squat Exercises, Journal of Strength & Conditioning Research — https://journals.lww.com/nsca-jscr/fulltext/2020/09000/relationship_between_velocity_loss_and_repetitions.18.aspx
• Refalo, Helms, Hamilton & Fyfe (2023), Influence of Resistance Training Proximity-to-Failure, Determined by Repetitions-in-Reserve, on Neuromuscular Fatigue in Resistance-Trained Males and Females, Sports Medicine - Open — https://pmc.ncbi.nlm.nih.gov/articles/PMC9908800/
• Armes, Standish-Hunt, Androulakis-Korakakis et al. (2020), Just One More Rep – Ability to Predict Proximity to Task Failure in Resistance Trained Persons, Frontiers in Psychology — https://pmc.ncbi.nlm.nih.gov/articles/PMC7785525/