Do You Need to Train to Failure to Build Muscle?

No. Stopping a rep or two short of failure builds about as much muscle and strength as grinding out every possible rep, according to a 2022 meta-analysis of 15 studies. What matters is that your sets are genuinely hard. For most lifts, end each set with about 1 to 3 reps left in the tank, and save true failure for light weights, machines and isolation exercises.

The short answer: get close, but you don't have to go all the way

Across 15 studies, training to failure and stopping short of it produced similar gains in strength and muscle size.¹ A few points matter more than failure itself:

• Effort still counts. "Stopping short" means 1 to 3 reps short, not 6. Growth seems to improve as sets end closer to failure.
• Light weights are the exception. With very light loads, a set only works well when it gets very close to failure.
• Failure has a cost. It brings more fatigue and slower recovery, and form tends to slip on the last reps.

What does "training to failure" mean?

Failure is the point in a set where you can't finish the lifting part of another rep, even though you're giving it everything.² On a bench press, it's the rep that stalls halfway up and needs a spotter.

That sounds clear, but studies don't all define it the same way. Some use this strict meaning, while others use looser definitions of when a set counts as "to failure." That's part of why results can differ between studies.³

What are reps in reserve (RIR)?

Reps in reserve is the number of extra reps you could have done when you stopped the set. If you finish a set of 10 and could have managed 12, you stopped with 2 reps in reserve, or "2 RIR." Zero RIR means you went to failure.

Coaches often pair it with a 10-point effort scale: a 10 means no reps left, a 9 means one left, an 8 means two left, and so on. A 2016 study of 29 people squatting proposed this RIR-based scale and found it a practical way to gauge effort from day to day.⁴

The catch is that people aren't great at guessing. A 2022 analysis of 12 studies (414 people) found lifters underestimated how many reps they had left by about one rep on average, with wide variation.⁵ Guesses got more accurate closer to failure and on sets of 12 reps or fewer. Training experience didn't seem to make people more accurate.

In practice, your "2 reps left" may really be 3. That's fine, and it means you can often push a little harder than feels natural.

What does the research say?

Failure vs stopping short: similar gains

The 2022 meta-analysis by Grgic and colleagues pooled 15 studies.¹ The strength analysis covered 394 people. The muscle-size analysis covered 7 studies and 219 people. Everyone was a young adult, and programs ran 6 to 14 weeks.

Neither approach won for strength or muscle growth. Two details are worth knowing:

• When the non-failure groups did more total sets, they actually gained more strength. When the amount of work was matched, the two approaches were almost identical.
• In trained lifters, failure showed a small edge for muscle growth. That result came from only 2 studies, and the authors called it preliminary.

The muscle-size estimate also had a wide margin of error, so a meaningful benefit from failure can't be ruled out yet. More research in older adults and highly trained lifters is still needed.

Closer to failure may help growth, up to a point

A 2023 meta-analysis of 15 studies found a trivial advantage for growth when sets ended at any kind of failure.³ When the authors looked only at true failure versus stopping short, the advantage disappeared. Their reading: the link between effort and growth probably isn't a straight line.

A 2024 analysis estimated reps in reserve for each study and modeled the trend.⁶ Strength gains looked similar across a wide range of reps in reserve. Muscle growth improved as sets ended closer to failure. The authors stressed that this was exploratory, because RIR had to be estimated from how each study described its training.

A direct test in trained lifters

In a 2024 trial, 18 trained lifters (12 men and 6 women) trained one leg to failure and the other leg to about 2 reps in reserve on the leg press and 1 on the leg extension, twice a week for 8 weeks.⁷ Quadriceps growth was nearly identical: 0.181 cm vs 0.182 cm of added thickness. The failure leg lost more reps and bar speed within each session. It's a small study of two leg exercises, but it fits the bigger picture.

The ACSM's 2026 position stand, built from 137 systematic reviews, reached the same conclusion. Sets taken to failure didn't improve strength, muscle growth or power, and finishing near failure, about 2 to 3 reps in reserve, provides enough effort.²

Does the weight you use change the answer?

Yes. How close you get to failure matters more as the weight gets lighter.

A 2017 meta-analysis of 21 studies compared light loads (60% of your max or less) with heavier loads, with every set taken to failure.⁸ Muscle growth was similar. Heavier loads built more one-rep-max strength.

A 2022 study shows why the "to failure" part matters with light weights.⁹ Twenty-five untrained men trained each leg on the knee extension for 8 weeks, with total work matched across conditions:

• At 80% of max, quad growth was about the same whether sets went to failure (8.1%) or stopped short (7.7%).
• At 30% of max, going to failure grew the quads 7.8%. Stopping at about 20 reps instead of about 34 gave just 2.8%, too small to be statistically significant.

So with light weights or very high reps, take sets close to failure. With heavier weights, you have more leeway. It was one exercise in untrained men, so the exact numbers won't transfer to every lift.

Why not just go to failure on every set?

Because it tends to cost more than it gives.

• Slower recovery. In a study of 10 trained men, three sets of 10 to failure on the bench press and squat caused bigger drops in jump height and bar speed than the same total reps done as six sets of 5. Signs of fatigue lasted longer too, up to 24 to 48 hours.¹⁰
• Fewer reps on later sets. Failure sets drain the sets that follow. In the 2024 trial, the failure leg lost more reps from set to set.⁷
• Form breaks down. The last grinding rep of a heavy squat or bench press is where technique slips. The ACSM authors suggest failure may be a poor fit for some groups, such as older lifters, partly because poor form raises the chance of injury.²

How close to failure should you train?

Here's a practical starting point. It's guidance built from the studies above, not a tested protocol.

How to start using reps in reserve

1. Pick a rep range and a target, for example 8 to 10 reps with 2 in reserve.
2. After each set, ask yourself: how many more clean reps could I have done? Note the answer.
3. Check your guesses now and then. Take a safe exercise, like a machine or a curl, to true failure and compare the reps you got with what you expected.
4. Add weight when the target gets easy. ACSM guidance suggests adding 2% to 10% once you can do one or two reps more than your target.¹¹
5. Log every set so you can see when reps stall. Rep Labs plans your reps and weight from your logged history, with plateau detection and smart deloads.

When does going to failure make sense?

• On light-load or high-rep sets, where getting close to failure is what makes the set work.
• On the last set of a machine or isolation exercise, as a check on your effort.
• Occasionally, for trained lifters chasing extra growth, though the evidence for a benefit rests on just 2 studies.
• Not on heavy barbell lifts without a spotter or safety bars.

Effort is only one part of a set. How long you rest matters too, so see How long should you rest between sets? And if you're deciding between barbells and machines, see Free weights vs machines.