Proximity to Failure by Movement Type: Compound Squats vs Isolation Lateral Raises

No, squats and lateral raises should not end at the same distance from failure. For heavy compound lifts like squats, deadlifts and barbell rows, stopping around 1-3 reps in reserve (RIR) is the safer, more sustainable default. For simple single-joint lifts like lateral raises, curls and pushdowns, taking the last sets to 0-1 RIR costs far less.

That is the working answer to failure proximity compounds vs isolations. It sits inside the bigger picture in our guide to load, rep ranges, RIR and rest intervals, and below we look at what the research does and does not say.

Is there a different failure rule for compounds and isolations?

Be clear about one thing first. No trial has randomised lifters to “squats to failure, lateral raises to failure” versus the reverse and measured the difference. The rule above is a coaching heuristic, built from three solid pieces of evidence and a bit of common sense about risk.

• Failure does not clearly build more muscle.
• Failure makes you recover more slowly.
• A failed rep is far more dangerous with a heavy bar on your back than with an 8 kg dumbbell in your hand.

Put those together and the movement-type rule almost writes itself. Where the cost of failure is high, stay back. Where it is low, push.

Does training to failure build more muscle?

Less than gym culture suggests. A 2023 systematic review and meta-analysis of 15 studies found a trivial advantage for training to “set failure” overall (effect size 0.19), but no advantage for momentary muscular failure versus non-failure (effect size 0.12, confidence interval -0.13 to 0.37) (Refalo et al., 2023). The authors concluded there was no evidence that momentary failure is superior for hypertrophy.

A 2024 meta-regression took a different approach. It treated proximity to failure as a continuous number (estimated RIR) and found that hypertrophy improved as sets ended closer to failure, while strength gains were similar across a wide range of RIR (Robinson et al., 2024). The authors flagged that the RIR values were estimated from study descriptions and called the analysis exploratory.

So do these two conflict? Not really. The first asks “failure or not?”, the second asks “how close?”. A reasonable reading is that closer is somewhat better, but the gain flattens near the end, so the last rep or two probably adds little. That is why 1-3 RIR on hard compounds likely gives up only a small amount of growth.

What does failure cost you in fatigue and recovery?

This is where compounds differ most. In a crossover study, ten resistance-trained men did three protocols on bench press and full squat, including 3 sets of 10 reps with 10 possible (failure) versus 6 sets of 5 reps with 10 possible (same total volume, no failure). Failure sets caused a larger acute drop in jump height and bar velocity, and recovery between 24 and 48 hours was slower (Morán-Navarro et al., 2017). The authors suggested that avoiding failure leaves you in better condition for the next session.

Squat-specific data point the same way. In 15 resistance-trained men doing three sets of full squats at 70% of 1RM, protocols that allowed more velocity loss within the set (30% and 40%) produced more fatigue and slower recovery than a 20% velocity-loss protocol, with performance dips still showing at 24-48 hours (Cornejo-Daza et al., 2024). Velocity loss is a proxy for how close you got to failure, so this fits the same story.

A small 2026 crossover of six trained men found a full-body session to failure raised markers of muscle damage and soreness and delayed jump recovery through 72 hours versus a non-failure session with similar total volume (Vieira et al., 2026). Six participants is tiny, so the authors themselves urged cautious interpretation, but it matches the larger-sample pattern.

Notice what most of these studies used: squats and bench press. Large-muscle, heavy lifts. That is exactly where fatigue costs pile up.

Why compounds need a bigger buffer

Fatigue is only half of it. The other half is mechanics.

On a heavy squat, a bad last rep means your hips shoot up first, your lower back rounds, or the bar drifts forward. All of that happens while the bar sits on your upper back. A deadlift or barbell row asks your back to hold a position under load while your legs and grip also tire.

In gym terms, this looks like the “last rep” that looks nothing like your first rep. If you could not film the rep without wincing, it was one too many.

Compounds also use a lot of muscle at once, so they are tiring for the whole body and slow your recovery for the next session. For a deeper look at that, read our piece on how central and peripheral fatigue limit lifting.

Why isolations can go to failure with a smaller penalty

Take lateral raises. The load is light, the joint is simple, and the muscle that fails (the lateral deltoid) is small. You are not balancing a heavy bar and your spine is not under axial load. If a rep stalls, you just lower the dumbbells.

Evidence supports using failure on isolations. A 2025 trial had 24 resistance-trained men and women train lateral raises for five sets to momentary failure, twice a week for eight weeks. Lateral deltoid thickness rose by 3.3-4.6%, with similar results from dumbbells and cables (Larsen et al., 2025). That shows failure on lateral raises is workable and productive, not that it beat stopping short, since the study did not include a non-failure group.

The fatigue cost is also smaller because less total muscle is involved. Raising 7 kg for the last three reps of a set does not leave you tired for two days the way grinding squats can.

For setup and technique, see our lateral deltoid guide on dumbbells versus cables and machines.

So what does a smart mix look like?

Think of failure as a tool you spend, not a badge.

• Early sets: leave 2-3 RIR on every lift.
• Last set of an isolation: go to 0-1 RIR.
• Last set of a compound: 1-2 RIR, rarely lower.
• Machine compounds (leg press, chest-supported rows): somewhere in between, 0-2 RIR.

This idea fits a broader principle: choose exercises and effort so you get the most stimulus per unit of fatigue. We cover that in stimulus-to-fatigue ratio and exercise selection. If you want the meta-analysis story in more depth, see training to failure versus RIR for hypertrophy.

How to decide set by set

Use three questions before pushing a set past about 1 RIR.

1. Does a bad rep put your spine, a barbell or a joint at risk?
2. Does your form stay the same on the final reps?
3. Will you recover before the next session that uses this muscle?

If the answer to the first is yes, stop at 1-3 RIR. If your form changes, end the set a rep or two earlier. Only when all three are clear should you go to 0.

You can also set loads by effort with our RIR and RPE load calculator.

Can you actually judge failure accurately?

Reasonably, but not precisely. On average, lifters tend to underpredict how many reps they have left, usually by a rep or two, and guesses are less accurate far from failure. That is why the 1-3 RIR zone is forgiving: even a guess that is off by one rep keeps you out of real trouble. We break down the studies in how accurately lifters gauge RIR, and our RIR and RPE scale guide covers the scale itself.

What about velocity loss and bar speed?

If you have a way to track bar speed, velocity loss is a more objective stop signal than “feels hard”. In 64 resistance-trained men squatting for 8 weeks, groups stopping at 20% and 40% velocity loss showed significant hypertrophy increases, but only the 40% group showed negative neuromuscular changes such as slower early rate of force development (Pareja-Blanco et al., 2020). The authors suggested moderate thresholds for strength and to avoid negative adaptations. Most lifters do not own a velocity tracker, but the lesson transfers: when the bar slows a lot, the set has done its work.

Who should be extra careful?

Use extra margin if you are new to a lift, returning from injury, training alone without safety pins or a spotter, dieting hard, sleeping badly or training in heavy heat. In Indian summer conditions, with long commutes and short sleep, recovery is often the thing that runs out first. Cut your effort a notch before cutting your frequency.

Sharp pain, pinching or joint symptoms are a stop sign at any RIR. If pain lingers beyond a day or two, or you feel numbness or pain shooting down a limb, see a doctor or physiotherapist rather than training through it.

The bottom line

Squats and lateral raises should not share one failure rule. Leave 1-3 reps in reserve on heavy compound lifts, let simple isolations reach 0-1 RIR on the last set, and treat failure as occasional seasoning rather than the main meal. The studies behind this show that failure adds little growth but real fatigue, and that the practical risk lies mostly in the big lifts.

Want a structure that puts this into practice? Try our free leg day plan, and see our guide to junk volume and diminishing returns if your sessions feel endless.