The Case for Training to Failure (Sometimes)

Most training advice lives in absolutes. Either you should always train to failure because anything less is lazy, or you should never train to failure because it's inefficient and wrecks recovery. Both camps are wrong.

The truth is more useful and more boring: training to failure has a place, but that place is smaller than the hardcore crowd thinks and larger than the auto-regulation purists admit.

What the Research Actually Shows

Studies comparing training to failure versus stopping short of failure generally show similar hypertrophy outcomes when volume is equated. The key phrase is "when volume is equated." If you do three sets of eight reps to failure, you've done roughly the same effective work as someone who does four or five sets of eight at RPE 7-8.

The practical problem: most people don't actually equate volume. They do three hard sets and call it done, when they needed five moderate sets to accumulate the same stimulus. Training closer to failure is a volume-efficient strategy, not a magic intensity threshold.

For strength development, the picture shifts. Research on proximity to failure suggests that staying 2-4 reps from failure on compound lifts produces similar strength gains to going all the way, especially for trained lifters. The nervous system adapts to heavy loads whether or not you grind out that last impossible rep.

Where Failure Makes Sense

Isolation work. If you're doing lateral raises or leg curls, going to failure occasionally gives you clear feedback about your effort level. These movements don't create the same systemic fatigue as a heavy squat, and the injury risk from one more rep of cable flies is essentially zero.

Last set of a hypertrophy block. Pushing that final set to true failure can squeeze out a bit more stimulus without dramatically increasing your weekly fatigue load. One set to failure out of fifteen total sets for a muscle group isn't going to wreck your recovery.

Testing your RPE calibration. Most people are terrible at judging proximity to failure early in their training career. Occasionally taking a set to failure teaches you what RPE 9 actually feels like, which makes your submaximal training more accurate. You can't get better at leaving two reps in the tank if you don't know where the tank ends.

Deloads and tapers paradoxically. When you're reducing volume before a competition or testing week, taking your reduced sets closer to failure can maintain intensity while dropping overall fatigue. Two sets to failure beats five moderate sets for peaking purposes.

Where Failure Is Stupid

Compound movements under fatigue. The last rep of a fatigued deadlift set is where form breaks down and backs get tweaked. The risk-reward calculus doesn't favor grinding out that final rep when you're already wobbling.

Every single set. Training to failure on every set of every exercise creates a recovery debt you can't repay with anything short of pharmaceutical assistance. The fatigue accumulates faster than the adaptation. You'll feel like hell and your performance will crater within weeks.

When you're learning a movement. Technical failure should come before muscular failure when you're building motor patterns. That last ugly rep of a new squat variation isn't building strength, it's programming dysfunction.

During a volume accumulation phase. If you're running a high-volume block with multiple sessions per week, going to failure regularly will compromise the next workout. Leaving a couple reps in reserve lets you actually complete the program as written.

RPE Is a Tool, Not a Religion

The Rating of Perceived Exertion scale gives you a common language for intensity, but it's not a replacement for thinking. An RPE 8 on Monday after a weekend of sleep might be two reps from failure. That same RPE 8 on Friday after four hard training days might be four reps from failure because your CNS is fried.

RPE works best when you're honest about contextual fatigue. A true RPE 8 means you could definitely do two more reps with good form right now. Not two more reps if you rested five minutes. Not two more reps on a fresh day. Right now.

Most intermediate lifters underestimate their proximity to failure by one to two reps. They think they have three left when they have one. This is where occasional failure testing helps. Once every few weeks, take a set you thought was RPE 8 and find out if you actually had two more reps or five more reps.

The Practical Middle Ground

Here's a framework that matches what the literature suggests and what actually works in practice:

Most of your compound lifting should live at RPE 7-8. You could do two to three more reps with solid form. This is hard enough to drive adaptation but conservative enough to sustain week after week.

Isolation and assistance work can venture to RPE 9 more frequently. One rep left in the tank on your last set of curls or rows won't destroy you.

Take something to actual failure once every week or two, usually on a less systemically fatiguing movement. This keeps your RPE calibration honest.

Never train to failure on movements where form breakdown creates injury risk under load. No failure deadlifts when you're tired. No failure overhead press when your shoulder is cranky.

This Week

Pick one isolation exercise in your next workout. On the last set only, go to actual muscular failure where you physically cannot complete another rep. Notice how many reps that was beyond what you normally do. That's your RPE calibration error.

For your main lifts this week, aim for RPE 8 and film a set. Watch it back. Could you have done two more reps with that same bar speed and form? If the answer is obviously yes, you're training softer than you think. If the answer is obviously no, you're closer to failure than your recovery can sustain long-term.

The goal isn't to always train to failure or never train to failure. The goal is to accumulate enough hard sets to grow while staying healthy enough to train again in two days. Sometimes that means stopping at RPE 7. Sometimes that means grinding out a final rep. The difference is knowing which situation you're in.