How close to failure should you train? 1 to 3 reps short
Close to failure means stopping with 1 to 3 good reps left. You keep the strength and nearly all the muscle for less fatigue. How to tell where you are.
Updated

In this article8
The short answer
Close to failure means ending a set when you could do 1 to 3 more good reps. You get the same strength and nearly the same muscle as going all the way, for far less fatigue. Most people have more reps left than they think.
Failure is the rep you start and can't finish. Not the rep that burns, and not the one where you'd quite like to stop. It's the one where the weight slows to a halt halfway up, however hard you push. (Some coaches use a softer line: the moment your form starts to slip. Either way, close means stopping before you get there, with clean reps still in hand.)
So the number is easy. The hard part is knowing, in the middle of a set, whether you're 2 reps from that point or 7.
What 2 reps left looks like
Picture a set of one-arm dumbbell rows, with a weight you could lift 11 times before the 12th gets stuck. This is roughly how long each rep takes on the way up.
An illustration of a typical set, not measured data
Reps 1 to 7 feel much the same. You could hold a conversation. Around rep 8 the dumbbell suddenly feels heavier than it did a moment ago, though it still moves at a fair pace. That's 3 left. By reps 9 and 10 the pull is visibly slower, you want a breath at the top before the next one, and your body starts offering shortcuts: a twist of the torso, a little jerk to get the weight moving. Rep 11 is a grind the whole way up. Rep 12 stops halfway.
So in this set, stopping after rep 8, 9 or 10 is training close to failure. Rep 11 is failure in all but name.
The signs you're in that zone, in order of how much to trust them:
- The rep is slower than your first ones, even though you're pushing just as hard. It's the signal to watch first, and it's real: in lab tests, slower reps go with higher effort ratings.1
- You want a pause between reps: a rest at the top, a big breath before the next.
- Your form wants to change. An elbow drifts out on a press, your back arches, your hips rise first on a squat, you lean back to finish a curl.
- You can tell the next rep would be a fight, before you start it.
Burning and shaking are less useful. On a long set with a light weight, the burn can arrive well before you're anywhere near the end.
If your program or app says "RIR 2" or "RPE 8", it's asking for the same thing: stop with 2 good reps left. RPE 9 means 1 left, and RPE 10 means you went to failure.1 There's more on both scales in rating effort like a coach.
Do you lose anything by stopping short?
For strength, no. Trials that compared the two found people got just as strong stopping short as going to failure,23 and the American College of Sports Medicine's 2026 guidelines name stopping with 2 or 3 reps left as one way to train hard enough.4
For muscle, close does matter a little. Growth creeps up as sets end nearer failure, though the researchers who found that trend call it exploratory.3 But the very last rep adds the least. When experienced lifters trained one leg to failure and the other stopping 1 or 2 reps short, both thighs grew by the same amount over eight weeks.5
What going all the way does buy you is tiredness. In one bench press experiment, going to failure left lifters about twice as worn out straight afterwards as stopping 1 rep short. It felt worse and left them sorer, and it didn't even add up to more reps, because the all-out early sets cost reps on the later ones. Two days later the bar was moving at full speed again, though they still felt a little sorer and less recovered.6 Not a lasting harm, then, just a tax you pay every time for very little back.
Most people stop too early
In one lab test, experienced lifters doing a long set of squats called out when they thought they had 5, 3 and 1 reps left, then kept going until they couldn't.7
Zourdos and colleagues, 2021
Even at "1 left", the call was off by about 2 reps. Averaged across studies, the miss runs one way, toward reps still in the tank, and it grows the further from failure you are and the longer the set. Years of training didn't seem to fix it.78
So until you've tested yourself, add a rep or two to whatever it feels like, and make the call on the last few reps, not at the start, because the closer you are the better you guess. That matters most with light weights.
If your dumbbells are light, go closer
Light weights build muscle too, but they leave the most room for a bad guess. A set can feel tough at rep 20 and still be a long way from the end.
One trial on leg extensions shows what that can cost. With heavy weights, stopping short grew muscle as well as going to failure. With light weights, the leg taken to failure grew just as much as the heavy ones, but the leg that stopped around rep 20, with about 15 still in it, barely grew.9
That's one small trial in untrained men, so treat it as a warning rather than a law. But it fits the guessing research: fifteen reps short sounds like a lot, but with light weights it's easy to do without noticing.
If your sets regularly run past 20 reps, you have three good options:
- Push them closer, to within 1 or 2 reps of the end.
- Slow the reps down, three seconds on the way down, or add a pause, so the same weight runs out in fewer reps. Progressive overload at home without adding weight has more ways to do this.
- Get a heavier pair so most sets end somewhere between 8 and 12 reps, where your guesses are better. How much weight to start with has a simple test.
Find your own gap in one workout
The best fix for bad guessing is one honest set. It takes about five minutes.
Pick a move you can safely fail at alone
A dumbbell curl, a lateral raise, a machine, or push-ups against a kitchen counter or from your knees. Not barbell squats, deadlifts or bench press, where a stuck rep can trap you under the weight.
Do your usual set and stop where you normally would
Note the rep number, and make an honest guess: how many more could you have done?
Then keep going until a rep won't finish
Don't put the weight down. Same pace, same form. Count every clean rep, and stop the moment one stalls or your form breaks.
Work out your gap
Say you stopped at 12, guessed 2 left, and kept going to 17. You really had 5 left, so you were 3 reps short in your guess.
Use the answer next time
With that weight, stopping at 15 leaves you about 2 in the tank. Your old stopping point was 3 reps early.
Check again every few weeks
And whenever you start a new exercise. You get stronger, so the answer moves.
Filming yourself side-on helps, here and on any set, because the slowdown is easier to see on a replay than to feel mid-set (why form breaks under fatigue shows what to look for). It's also what a Bespoke Fit coach does on every live video session: they watch the weight slow and tell you "two more, then stop" while you're still holding it, so the call isn't left to your guess.
How close to go on each exercise
What decides the margin is what happens if you do miss a rep. Researchers who reviewed this suggest making that call mainly on safety.10 The table is our coaching guidance built on that, not trial data. The same margins apply to heavy sets of 3 to 6 reps.
| Exercise | Stop with | Ever to failure? |
|---|---|---|
| Barbell squats and deadlifts, heavy lunges and step-ups | 2 to 3 reps left | No. Form goes before the muscle does, and a missed rep under a heavy load is hard to get out of |
| Light goblet squats, bodyweight squats, lunges and step-ups | 1 to 2 left | Not needed. Slow the reps down instead, because a shaky last lunge or step-up can cost you your balance |
| Barbell bench press and overhead press | 2 to 3 left | Only with safety bars or a spotter |
| Dumbbell presses | 1 to 3 left | Rarely. A stuck dumbbell press is awkward to get out of |
| Dumbbell rows | 1 to 2 left | Now and then, on the last set |
| Machines and cables | 1 to 2 left | Yes, on the last set, with any safety stops set |
| Curls, lateral raises, triceps work, calf raises | 0 to 2 left | Yes, on the last set |
| Push-ups and glute bridges | 1 to 2 left | Yes, on the last set. You can stop anywhere |
The bigger, heavier and more complicated the move, the more room you leave. The smaller and safer it is, the closer you can go.
Two more things about failure. It's worth more when you only do a few sets for a muscle, or train it just once or twice a week.10 And it belongs at the end of an exercise, not the start: an all-out first set drags on every set after it, which is what the bench press experiment above showed.
New, coming back, or over 50
Almost all of this research was done on young adults.2 The few trials in older adults point the same way: going to failure added nothing to strength, and where sets stopped well short, adding sets to make up the work built the same muscle.1112
So stop with 2 to 3 reps left. That's plenty while you learn how each move should feel, because a lot of your early progress comes from getting better at the movement itself, and grinding out ugly reps practises the wrong version. If you want more from a workout, add a set before you add effort. How many sets per week your muscles need gives you the numbers. The one exception is light dumbbells: if your sets run past 20 reps, finish them within a rep or two of the end, as in the light-weights section above.
Starting to lift at 50 with zero gym background walks through the first month. And if you've heard that older lifters need much longer to recover, the evidence is less clear-cut than you'd think.
If you have a heart condition or blood pressure that isn't under control, are pregnant, or are recovering from surgery, ask your doctor how hard you should push before taking sets near failure.
Common questions
Is training to failure bad for you? Not by itself, but it's rarely worth it. Going all the way mostly buys extra tiredness for little extra muscle. It's fine now and then on curls, machines and push-ups, where a failed rep just stops; on heavy barbell lifts, keep 2 to 3 reps back.
Should every set be close to failure? No. Warm-up sets and sets where you're learning a move should feel easy. The sets that count, the ones you write down, are the ones that should end within a few reps of failure. If those stop with 6 or 8 reps left, they do less for you than they could.
Does being sore the next day mean I went close enough? No. Soreness mostly tells you how new or unusual the work was, which is why it fades after the first few weeks of a program even when the sets are just as hard. Judge the set while you're doing it, using the signs near the top of this page.
How do I keep making progress if I always stop short? Keep the same reps-left target and let the reps climb. If you stopped at 10 with 2 left last week, and this week you can do 11 with 2 left, you got stronger. When the reps get high, move up a weight and start again a few reps lower.
Sources
-
Zourdos MC, Klemp A, Dolan C and colleagues, 2016. Novel resistance training-specific rating of perceived exertion scale measuring repetitions in reserve. Journal of Strength and Conditioning Research. ↩ ↩2
-
Grgic J, Schoenfeld BJ, Orazem J, Sabol F, 2022. Effects of resistance training performed to repetition failure or non-failure on muscular strength and hypertrophy: a systematic review and meta-analysis. Journal of Sport and Health Science. ↩ ↩2
-
Robinson ZP, Pelland JC, Remmert JF and colleagues, 2024. Exploring the dose-response relationship between estimated resistance training proximity to failure, strength gain, and muscle hypertrophy: a series of meta-regressions. Sports Medicine. ↩ ↩2
-
Currier BS, D'Souza AC, Phillips SM and colleagues, 2026. Resistance training prescription for muscle function, hypertrophy, and physical performance in healthy adults: an overview of reviews. Medicine and Science in Sports and Exercise. ↩
-
Refalo MC, Helms ER, Robinson ZP, Hamilton DL, Fyfe JJ, 2024. Similar muscle hypertrophy following eight weeks of resistance training to momentary muscular failure or with repetitions-in-reserve in resistance-trained individuals. Journal of Sports Sciences. ↩
-
Refalo MC, Helms ER, Hamilton DL, Fyfe JJ, 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. ↩
-
Zourdos MC, Goldsmith JA, Helms ER and colleagues, 2021. Proximity to failure and total repetitions performed in a set influences accuracy of intraset repetitions in reserve-based rating of perceived exertion. Journal of Strength and Conditioning Research. ↩ ↩2
-
Halperin I, Malleron T, Har-Nir I and colleagues, 2022. Accuracy in predicting repetitions to task failure in resistance exercise: a scoping review and exploratory meta-analysis. Sports Medicine. ↩
-
Lasevicius T, Schoenfeld BJ, Silva-Batista C and colleagues, 2022. Muscle failure promotes greater muscle hypertrophy in low-load but not in high-load resistance training. Journal of Strength and Conditioning Research. ↩
-
Refalo MC, Helms ER, Trexler ET, Hamilton DL, Fyfe JJ, 2023. Influence of resistance training proximity-to-failure on skeletal muscle hypertrophy: a systematic review with meta-analysis. Sports Medicine. ↩ ↩2
-
da Silva LXN, Teodoro JL, Menger E and colleagues, 2018. Repetitions to failure versus not to failure during concurrent training in healthy elderly men: a randomized clinical trial. Experimental Gerontology. ↩
-
Bergamasco JGA, da Silva DG, Bittencourt DF and colleagues, 2022. Low-load resistance training performed to muscle failure or near muscle failure does not promote additional gains on muscle strength, hypertrophy, and functional performance of older adults. Journal of Strength and Conditioning Research. ↩


