The lifting section made the case that most people stop sets earlier than they think. This piece makes what looks like the opposite case, and isn't: many of the reps performed after a certain point are no longer the exercise you started.
Both are true simultaneously. That's what makes proximity to failure genuinely difficult rather than just a matter of being tougher.
Coordination fails before force does
A multi-joint lift is a coordination problem before it's a strength problem. Producing force is the easy part; producing it in the right sequence, in the right direction, with the right joints stabilised, is the hard part — and it's the part fatigue attacks first.
Several things degrade in parallel as a set progresses:
- Neural drive to the prime movers falls. Central fatigue reduces the command reaching the muscles doing the primary work.
- Rate of force development declines faster than peak force. You can still generate force; you generate it more slowly. Any part of the lift depending on speed suffers disproportionately.
- Proprioceptive acuity drops. Joint position sense measurably worsens with fatigue, which means your internal model of where your body is becomes less accurate exactly when precision matters most.
- Movement variability rises. Bar path becomes less repeatable, timing between joints loosens, left–right asymmetries emerge or amplify.
Notice that none of these is "the muscle stopped working." They're all control problems, and they arrive before the muscle is anywhere near done.
The nervous system protects the task, not the target
Here's the key idea, and it's the one worth taking away.
Your motor system's objective is to complete the task — move the load from A to B. It has no representation whatsoever of your intention to train a particular muscle. That goal exists only in your head, as a training concept.
So when the prime mover fatigues, the system does the sensible thing: it recruits whatever else can contribute. Synergists take a larger share. Postural muscles start generating movement rather than stabilising it. Joint angles shift to positions with better leverage. Momentum substitutes for force. The task keeps getting completed, which from the motor system's point of view is a total success.
This is what "form breakdown" is, mechanically. It isn't sloppiness or lack of focus. It's an adaptive redistribution of work away from the fatiguing muscle — the precise muscle you were trying to fatigue.
What it looks like, lift by lift
The compensations are predictable enough to watch for:
- Range of motion shrinks silently. The most common and least noticed. Squat depth creeps up, the bar stops short of the chest, the pulldown stops short of full stretch. Because the failure is gradual and the count is unaffected, almost nobody perceives it — which matters enormously given how much range determines the stimulus.
- Trunk position changes. Increasing forward lean in a squat shifts load from quadriceps toward hips and back. It's a genuinely effective strategy for completing the rep and a complete failure at training the quads.
- Momentum appears. Hip drive in a curl, a leg kick in a row, a bounce out of the bottom. Reduces the force the target muscle must produce — which is the opposite of the goal.
- Tempo asymmetry. The eccentric gets dropped rather than controlled, cutting a substantial portion of the stimulus while making the set feel easier to continue.
- Asymmetry between sides. The stronger side takes over. Common on presses and rows, and effectively invisible without video.
Every one of these makes the rep easier for the target muscle. You experience continued struggle, because the effort is real — it's just being distributed to tissue you weren't trying to train.
The resolution: technical failure
So we have two errors pointing in opposite directions. Stop too early and the high-threshold units never get recruited. Continue too long and the reps stop loading the intended muscle while accumulating fatigue and injury risk.
The useful concept between them is technical failure: the last rep you can complete while the movement is still recognisably the same movement. Defined in advance, per exercise, with explicit criteria — a depth standard, a bar path, a tempo, a rule about momentum.
The productive region of a set sits just before technical failure. Those reps are slow, they're maximally recruited, and they're still delivering load to the target. Reps beyond it are largely a different exercise performed while tired.
What to do about it
Film sets. Actually film them. This is the whole intervention, honestly. Fatigue degrades your ability to perceive your own position at precisely the moment position is changing — you cannot audit this from the inside. One phone on the floor for one set per exercise will show you things you have confidently believed were not happening for years.
Write down the standard. "Squats to below parallel, no increase in forward lean, no bounce." A rep that misses the standard isn't a rep — it's the signal that the set is over. Without a written criterion, the standard drifts with fatigue, which is the failure mode itself.
Put failure where it's safe. Machines, cables and stable isolation work let you approach absolute failure with minimal technique cost and low risk. Heavy free-weight compounds should generally stop at technical failure. Matching the exercise to the intended proximity is most of the skill.
Use velocity loss as a trigger. If the bar has slowed markedly across the set despite maximal intent, you're in the productive region. If it's also started deviating from its usual path, you've passed the end of it.
Order the session accordingly. Coordination-heavy work goes first, while control is intact. Ending a session with technical barbell work means training a pattern under exactly the conditions that make it degrade.
Treat range as a hard boundary. A shorter rep is not a rep with a small penalty; it can be a fundamentally different stimulus. Better to end the set than to keep counting reps that no longer qualify.
The point
"Train close to failure" is good advice that quietly assumes something false — that the reps near failure are the same as the reps earlier in the set. Mechanically they often aren't, and the drift happens below the resolution of your own perception.
Which is the same structural problem as everywhere else in this archive. Your sense of effort misreports proximity to failure. Your sense of the muscle misreports tension. And your sense of your own position misreports what your body is doing under fatigue. In all three cases the fix is identical: get an external reference into the loop, because the internal one is not reliable at the exact moment it matters most.