One lifter slows the descent to a four-second count, holds at the bottom, and presses back up on a deliberate two-second tempo. The lifter next to them skips the counting entirely, adds another five kilos, and moves the weight as fast as they can control it. Both have built their programs around this difference — the first treating time under tension as the variable worth protecting, the second treating load as the only number that matters.
Each has seen results. Each has defended their approach in conversations that went longer than either planned. And each has had the same private moment after arguing with someone equally experienced who trains the opposite way — the brief pause where certainty flickers, and the question underneath it surfaces.
Does time under tension matter more than how heavy you lift
Three independent meta-analyses examined both sides of this debate. What they found did not come from the direction either camp expected.
Time under tension and load are both secondary variables for muscle growth. Three independent meta-analyses show that neither determines hypertrophy when training effort is sufficient. The actual driver is training volume — weekly sets taken near failure. The debate between time under tension and heavy weight asks the wrong question.
— Schoenfeld et al. 2017 · J Strength Cond Res · n=21 studies + Enes et al. 2025 · Scand J Med Sci Sports · n=14 RCTs
A 2017 meta-analysis pooled 21 studies comparing heavy loads against light loads for muscle growth. The result was not close. Heavy and light produced virtually identical muscle growth — differences so small that no statistical test could distinguish between them. The lifters who had been chasing progressive overload through weight were looking at evidence showing their most protected variable, across a wide range, did not determine how much muscle they built.
That should have been a win for the tempo camp. If load does not drive growth, maybe controlling rep speed does. A 2025 meta-analysis tested this directly — pooling 14 experiments where participants were randomly assigned to different rep speeds. Tempo had a trivial effect on muscle growth. The odds of it making even a modest difference were worse than a coin flip. The tempo camp's primary variable was not just secondary. It was negligible.
Neither camp was focused on what actually predicted growth. Muscle growth scales with training volume — weekly sets, performed with enough effort that the last few reps of each set are genuinely difficult. Volume had the strongest statistical support of any variable studied. Load and tempo both faded into noise once volume and effort were accounted for.
The effort threshold is not a footnote. The load findings apply when each set is pushed to the point where another rep is not possible. Without that intensity, heavier loads may carry a genuine advantage — a lifter who consistently stops well short of that point cannot assume their load selection is irrelevant. When volume is matched, the heavy-versus-light comparison draws this boundary more sharply.
The practical range for rep speed is wider than either camp assumes. Anything from about a quarter-second to four and a half seconds per action produces equivalent growth. Only extremely slow tempos — ten seconds or more per rep — show signs of falling behind. Within that window, the seconds a lifter counts on each rep are not driving their results.
Load is not entirely without value. For maximal strength, heavy training still produces clearly larger gains than lighter loads — even when muscle growth between the two is identical. The heavy camp has a legitimate case. Their mistake is extending it to hypertrophy, where the data does not support them.
Volume does not scale in a straight line, though. The returns flatten as weekly sets climb, which means the practical question is no longer which variable to optimize but how much volume is enough before more becomes wasted effort. For why the load question does not change the growth equation even at its extremes, the evidence on light versus heavy loads runs deeper than this comparison covers.