Training to failure means doing reps until you can't complete another one with good form. It's one of the most argued-about variables in lifting, and the evidence has moved a lot since most of the pages you'll find on it were written. The short version: going all the way to failure is not required to build muscle or strength, but getting close to it matters, especially for muscle growth. This guide shows you what the newest reviews of the evidence say, where failure helps, where it costs you, and exactly how close to push on each exercise.
What Training to Failure Actually Means
Coaches and scientists use "failure" in a few different ways, and the difference changes how you should read the evidence.
- Momentary muscular failure: you try the next rep and the weight doesn't move, or it stalls halfway. This is true failure.
- Technical failure: you could grind out another rep, but only by breaking form (bouncing, twisting, cutting range of motion). Many coaches stop sets here.
- Set failure or "close to failure": you stop with a set number of reps left in the tank. This is usually expressed as reps in reserve (RIR). A set at 2 RIR means you stopped when you could have done 2 more clean reps.
Most practical programming lives in that third category. When people ask "should I train to failure," the more useful question is "how many reps should I leave in the tank, and on which lifts?"
Does Training to Failure Build More Muscle?
Barely, if at all, when everything else is equal. Four pooled analyses of the controlled trials line up surprisingly well once you read the details.
A 2022 meta-analysis of 15 studies in young adults found no significant overall difference in muscle growth between training to failure and stopping short (effect size 0.22, with a confidence interval that crossed zero). In the subgroup of resistance-trained lifters, failure showed a small advantage for hypertrophy (effect size 0.15). The authors noted that all participants were young adults, so the results may not carry over to older or highly trained lifters.
An earlier meta-analysis of 13 studies found that failure training produced more hypertrophy overall (standardized mean difference 0.75), but that advantage disappeared when total training volume was matched between groups. In other words, failure seemed to help mainly because those groups ended up doing more work.
A 2023 meta-analysis of 15 studies on proximity to failure found only a trivial advantage for set failure over non-failure (effect size 0.19), and no advantage when it looked strictly at momentary muscular failure versus stopping short. The authors concluded that the relationship between effort and growth is probably not linear: getting close counts, but the last rep or two adds little.
The most recent piece is a 2024 series of meta-regressions that estimated the reps in reserve for every training group in the literature. Muscle growth increased as sets ended closer to failure. The catch: RIR had to be estimated from study descriptions, and the authors call the analysis exploratory. Treat it as a strong hint, not a precise dose chart.
Put together: sets that end very far from failure (say, 5 or more reps left) leave growth on the table. Sets that end within about 0 to 3 reps of failure all work well, and the difference between 1 RIR and true failure is small.
Does Training to Failure Build More Strength?
No. For strength, the evidence points the other way more often than not.
In the 2022 meta-analysis of 15 studies, there was no difference in strength gains between failure and non-failure training (effect size -0.09). When volume was not equated, non-failure training actually produced better strength gains (effect size -0.32). The 2021 analysis of 13 studies found the same pattern, plus better power output with non-failure training when volume was not matched.
The 2024 meta-regressions agree: strength gains were similar across a wide range of reps in reserve. Strength depends more on lifting heavy loads and practicing the movement with crisp technique than on how deep you dig on each set. Grinding every set to failure on your main lifts mostly adds fatigue, which cuts into how much quality volume you can do.
The Hidden Cost: Fatigue and Recovery
Failure sets are expensive. A 2017 study of 10 resistance-trained men compared 3 sets of 10 reps to failure against 6 sets of 5 reps with the same load (the same total reps, but stopping halfway through each set). On the bench press and full squat, the failure protocol caused a bigger drop in jump height and bar speed, and recovery took longer: the non-failure groups bounced back faster between 24 and 48 hours after training. The sample was small, but the finding matches what most lifters feel.
That recovery cost matters if you train a muscle two or three times a week. If failure on Monday leaves your squat flat on Wednesday, you've traded two good sessions for one hard one. You can read more about spacing your training in our guide to rest days and muscle growth.
When Training to Failure Makes Sense
Failure isn't bad. It's a tool that fits some situations better than others.
- Light loads. Sets of 20 to 30 reps can grow muscle about as well as heavy sets. A meta-analysis of 21 studies found similar hypertrophy between low-load (60% of 1RM or less) and high-load training, but every set in those studies was taken to failure. With light weights, the first 15 reps are easy, so you need to push close to failure to make the set count.
- Isolation and machine exercises. Curls, lateral raises, leg extensions, cable work and most machines are low risk to fail on and cause less whole-body fatigue than a heavy squat.
- The last set of an exercise. Taking only your final set to failure lets you check how accurate your RIR guesses are without paying the cost on every set.
- Short phases before a deload. A week or two of higher effort right before a lighter week is a classic way to push hard and then recover.
- Techniques built around failure. Methods like rest-pause training and drop sets use failure on purpose. Save them for the end of a session.
When to Stop Short of Failure
- Heavy compound lifts. Squats, deadlifts, bench presses and overhead presses are where form breaks down first and where a missed rep can hurt you. Leave 1 to 3 reps in reserve.
- Strength or power goals. If you're peaking a lift or training for a sport, stay further from failure (2 to 4 RIR) and spend your energy on heavy, fast, technically sharp reps.
- High-frequency programs. If you hit the same muscles three or more times a week, frequent failure sets make recovery harder.
- Beginners. New lifters grow from almost any reasonable effort, and they're still learning what a clean rep feels like. Technical failure is the right stopping point.
Safety note: never take a barbell squat or bench press to failure without a spotter or properly set safety pins. If you train alone, use machines, dumbbells you can drop safely, or stop at 1 to 2 RIR on barbell lifts. If you have a heart condition, high blood pressure or a recent injury, ask your doctor before doing all-out sets.
How Close to Failure Should You Train? A Practical Plan
Use this as a default and adjust based on recovery and progress.
Effort targets by exercise type
- Heavy compounds (squat, deadlift, bench, overhead press, rows): 3 to 5 sets of 4 to 8 reps at 1 to 3 RIR. For example, if you can squat 225 lb (102 kg) for 8 reps, do sets of 6.
- Secondary compounds (lunges, Romanian deadlifts, incline press, pull-ups, machine presses): 2 to 4 sets of 6 to 12 reps at 1 to 2 RIR.
- Isolation and machines (curls, extensions, raises, leg curls): 2 to 4 sets of 10 to 20 reps at 0 to 1 RIR. Taking the last set to true failure is fine here.
A sample upper-body session
- Bench press: 4 sets of 6 at 2 RIR (for example 185 lb or 84 kg if your 8-rep max is around there)
- Chest-supported row: 3 sets of 8 at 2 RIR
- Incline dumbbell press: 3 sets of 10 at 1 RIR
- Lat pulldown: 3 sets of 10 to 12 at 1 RIR
- Cable lateral raise: 3 sets of 15, last set to failure
- Dumbbell curl: 2 sets of 12, last set to failure
Rest long enough that you can hit your target reps on the next set. Our guide on rest between sets covers how long that usually is.
How to Gauge Reps in Reserve Accurately
Most lifters underestimate how many reps they have left, especially on higher-rep sets. A few ways to calibrate:
- Test it on a safe exercise. On a machine or isolation lift, pick a weight, guess your RIR at the end of a set, then keep going to failure. If you said 2 RIR and got 5 more reps, you've been stopping too early.
- Watch bar speed. Reps slow down noticeably as you approach failure. When a rep takes about twice as long as your first rep, you're usually within 1 to 2 reps of failure. Bar-speed tools make this objective; see our guide to velocity-based training.
- Track your sets. If you log weights and reps, a set at 2 RIR should become a set you can do for 2 more reps a few weeks later. That's progress you can see.
- Let effort guide the plan. Using RIR to adjust load day to day is the heart of autoregulation, and it keeps hard training sustainable.
Training to Failure FAQ
Does training to failure work?
It works, but it isn't required. A 2022 meta-analysis of 15 studies found that sets ending a few reps short of failure build muscle and strength about as well as failure sets when total volume is similar. Failure adds the most value on light loads and isolation exercises, and costs the most on heavy compound lifts.
Is 3 reps in reserve good?
Three reps in reserve is a solid target for heavy compound lifts and for strength-focused training. For muscle growth on isolation exercises, getting closer (0 to 2 RIR) is usually better, since the newest analyses suggest growth improves as sets end nearer to failure. Many lifters also underestimate their RIR, so your "3" may really be 4 or 5.
Does 3 RIR build muscle?
Yes. Sets ending around 3 reps from failure still produce meaningful growth, especially if you do enough weekly sets. The 2024 meta-regressions suggest slightly more growth as you get closer to failure, so if muscle size is your main goal, push your isolation work and last sets a little harder.
How do you calculate reps in reserve?
Reps in reserve is the number of additional clean reps you believe you could have done when you stopped the set. If you could have done 10 reps and stopped at 8, that's 2 RIR. To sharpen your estimate, occasionally take a safe exercise to failure after guessing, and note how far off you were.