Training to failure: how close is close enough

The bar stalls halfway up. You fight it for two seconds, it settles back onto your chest, and the set is over. Failure feels like proof you worked hard.

Usually it's proof you did two reps too many. The reps that build muscle are the hard ones just before failure, and by the time you're a rep or two out, you've already banked them. Going the rest of the way buys a lot of fatigue and almost no extra growth.

So the short version: one to three reps left in the tank on most working sets, two to three on heavy squats, deadlifts and presses, and true failure spent on purpose in a couple of specific places. The harder part is knowing where you actually are inside a set. Most people are worse at that than they think.

Two different things called failure

"Failure" sounds precise. In the research it means at least two things, and that blur is most of why the numbers look muddy.

Momentary muscular failure is the strict version. You try another rep and physically can't finish it through a full range of motion. Technical failure is softer: you stop when your form starts to go, even though an ugly rep or two is probably still in there. A trial that lets people grind out a shaky, half-locked rep is measuring a different event from one that stops at the last clean rep.

Reps in reserve is cleaner language. RIR is just how many good reps you had left when you racked the bar. Zero RIR is true failure. Two RIR means you stopped with two solid reps in the tank. I've written a full guide to rating effort like a coach, so I'll leave it there. Aiming at a number you can name beats chasing a stop line nobody defines the same way.

The last few reps do the building

Your muscle fibers get called up in order of size, smallest first.

Easy effort runs on small, fatigue-resistant fibers. They're good at going all day and they don't have much growth potential. As they tire, your nervous system reaches for bigger, higher-threshold fibers to keep the weight moving. Those bigger fibers are the ones with room to grow, and they only come fully online when a set gets genuinely hard, in the last handful of reps before you'd fail. Researchers call those the stimulating reps.

You don't have to fail to reach them. Two reps out, the big fibers are already working flat out and the bar is already slow. The stimulus is collected. What comes after is mostly fatigue piled on a job that's already done.

That's the part that feels wrong at first. There's no bonus for the final rep. Your muscle can't tell the difference between a rep you barely finished and one you barely missed.

What stopping short actually costs you

Very little, and the evidence here is unusually consistent.

A 2021 meta-analysis by Grgic and colleagues in the Journal of Sport and Health Science pooled 15 trials comparing training to failure against stopping short. Strength came out level, effect size -0.09. Muscle size showed a small edge for failure that wasn't statistically significant, 0.22, with the range crossing zero.¹ Another pooled analysis of 15 trials called the size advantage trivial: 0.19, and the bottom of that range sat right on zero.²

Go close and you keep essentially all of the strength and nearly all of the muscle. The last rep to a dead stop buys a sliver of size at most, and it isn't free.

The bill comes due two days later

In one bench press experiment, trained lifters did sets to failure, to 1 RIR and to 3 RIR while researchers tracked how fast their muscles came back. Four minutes after the set, bar speed was down 25% in the failure group, 13% at 1 RIR, and 8% at 3 RIR. A full day later, the failure and 1-RIR lifters were still slower than normal. The 3-RIR group had bounced back.³ Same lift, same load, wildly different recovery bills.

That fatigue stacks up. A set to failure damages more tissue and drains the nervous system harder, which drags on every set after it in the session and on the sessions that follow. On a heavy squat or deadlift the cost is highest: those lifts load the most muscle and the most spine, and a failed rep is exactly where technique comes apart.

Age sharpens the same effect. Recovery slows as you get older, so one failed set eats further into the week. Train closer to failure and you stay tired longer, and that tax is steepest on the big lifts and on older bodies.

Where failure earns its keep

None of this makes failure useless. It makes it something you spend deliberately instead of by accident on every set.

The first good use is the last set of an isolation move: a biceps curl, a lateral raise, a calf raise. One small muscle, safe to fail without a spotter, barely a dent in your recovery. There's some signal that small muscles reward going all the way. In a 2025 calf-raise experiment, lifters who pushed past failure with partial reps grew the calf about 9.6%, against 6.7% for stopping at failure.⁴ That's a calf, though, not a squat.

The second use is calibration. Most people misjudge how close to failure they are, and the fix is to take a set all the way now and then so you have a live reference for what zero RIR feels like. Do that occasionally and your everyday "two reps left" gets a lot more honest. It's the loop I walk through in the RIR guide.

So fail on small lifts, and fail to calibrate. Just don't let it become your default on everything else.

The trials were mostly run on young people

The research here is good, not airtight, and it's worth holding loosely in a few places.

Studies define failure differently, which widens the error bars on every comparison in those pooled analyses. And the size advantage for failure, small as it is, keeps landing on the same side of zero. Someone chasing maximum size on isolation work can reasonably push harder than someone lifting for health and longevity.

Then there's the sample problem. Nearly all of the failure-versus-stopping-short trials ran on young adults, so how cleanly this maps onto a 55-year-old is an open question, and the safer bet for older or newer lifters is more in reserve, not less. Novices are a case of their own: they tend to stop well short of true failure while feeling like they're at the limit, so a beginner's "failure" and a trained lifter's "failure" aren't the same event. Recovery also varies a lot from person to person. None of these averages tell you how your body will answer this week.

The rules I train by

• Most working sets stop at 1 to 3 reps in reserve. That collects the stimulating reps at a recovery cost you can pay again in two days.
• Heavy compounds, squats, deadlifts and presses, stay at 2 to 3 RIR almost always. Highest recovery bill, real form risk.
• True failure goes on the last set of isolation work you can fail safely alone: curls, raises, leg extensions.
• Over 50, or new to lifting? Leave more in reserve, not less. Recovery is your limiter, not effort.
• Every few weeks, take one set to genuine failure to recalibrate what zero RIR feels like, then go back to stopping short.

Common questions

Should you train to failure to build muscle? Usually not. Most of the stimulus comes from the last few hard reps before failure, and the pooled research finds only a small, often non-significant size advantage for going all the way. Stopping one to two reps short gets you nearly the same growth for a fraction of the fatigue.

Is training to failure bad for you? It's not dangerous by itself, it's expensive. Failing a heavy compound lift raises the odds of your form coming apart, and it slows your recovery enough to cost you quality sets later in the week. Used occasionally and on the right exercises it's a useful tool. As a daily habit it's a tax.

How do I know if I'm actually close to failure? Watch the bar. When a rep slows to a grind and your range of motion starts shrinking, you're within a rep or two. Most people, beginners especially, feel closer than they are, which is why taking the odd set to true failure keeps the rest of your estimates honest.

Does stopping short still make you stronger? Yes. Strength gains barely notice whether you reached failure. The 2021 meta-analysis found no meaningful strength difference at all, so you can leave reps in reserve and still add weight to the bar over time.

Reading your own effort from inside the set

One to three reps in reserve, logged honestly and pushed up over time, gets this job done on its own.

What's hard is the thing the calibration research keeps pointing at: judging your own effort while you're in the middle of the set. It's genuinely difficult to see your own bar decelerate, and easy to talk yourself into or out of how close that last rep really was. A coach watching live is an outside read on exactly that, calling it when your "two reps left" is really five, or waving you off a squat that's about to fold. That's why our coaches watch every rep as it happens instead of reading a log the next day, and it works alongside getting your weekly set count right. Founding membership is $49 your first month, then $299/mo locked for life, with 25 spots. A sharp training partner can do a fair bit of this for free. What you're paying for is an eye that reads effort you can't feel yet.

Sources

1. Grgic, Schoenfeld, Orazem, and Sabol, 2021. 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. Refalo, Helms, Trexler, Hamilton, and Fyfe, 2023. Influence of resistance training proximity-to-failure on skeletal muscle hypertrophy: a systematic review with meta-analysis. Sports Medicine. ↩

3. Refalo, Helms, Hamilton, and 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. ↩

4. Larsen, Swinton, Sandberg, Kristiansen, Fredriksen, Falch, van den Tillaar, and Wolf, 2025. Resistance training beyond momentary failure: the effects of past-failure partials on muscle hypertrophy in the gastrocnemius. Frontiers in Psychology. ↩