Time Under Tension, Training Frequency, and Proximity to Failure: What the Meta-Analyses Actually Show

Three of the most argued-about questions in strength training are how slowly you should lower the weight, how often you need to train a muscle each week, and how close to failure each set actually needs to go. All three have now been studied enough that meta-analyses — not single studies, not opinions — can answer them with real confidence. The answers are more forgiving than most gym folklore suggests. Time Under Tension: A Surprisingly Wide Range Works "Time under tension" (repetition duration — the combined time of the lowering, lifting, and any pause in a single rep) has long been treated as something to optimize carefully, with slow, controlled reps often assumed to be superior for muscle growth. According to PubMed, a systematic review and meta-analysis pooling eight controlled studies found that hypertrophic outcomes were similar across repetition durations ranging from 0.5 to 8 seconds per rep ([Schoenfeld, Ogborn & Krieger, Sports Medicine, 2015, PMID: 25601394](https://doi.org/10.1007/s40279-015-0304-0)). That's a genuinely wide range — from fast, explosive reps to fairly slow, deliberate ones — producing comparable muscle growth. The one caveat the same review flagged: very slow durations beyond about 10 seconds per rep appeared inferior for hypertrophy, though the authors noted there wasn't enough controlled research at those extremes to be fully confident in that conclusion. A more recent Bayesian meta-analysis specifically isolating tempo (rather than total rep duration) compared "slower" tempos (averaging roughly 3.5 seconds per rep) against "faster" tempos (averaging roughly 1 second per rep) and found a trivial difference between them once other training variables were controlled. The practical takeaway from both bodies of research points the same direction: mechanical tension from lifting a genuinely challenging load appears to be what drives growth, and the specific speed at which you do it matters far less than gym culture often implies. Training Frequency: It's About Volume, Not the Calendar The "how many days per week should I train chest" debate has a cleaner answer than the endless forum arguments suggest. According to PubMed, a systematic review and meta-analysis of 25 studies examining weekly training frequency found no significant difference in muscle growth between higher- and lower-frequency training when total weekly volume was equated ([Schoenfeld, Grgic & Krieger, Journal of Sports Sciences, 2018, PMID: 30558493](https://doi.org/10.1080/02640414.2018.1555906)). In other words, doing 12 total sets for a muscle group once a week produced similar growth to spreading those same 12 sets across two or three sessions. The review did find a modest advantage for higher frequency when volume wasn't equated between groups — largely because it's easier to accumulate more total volume across a week when it's spread across more sessions, rather than frequency itself being the active ingredient. The practical implication: frequency is a scheduling tool, not a growth mechanism in its own right. If you can fit your needed weekly volume into fewer, longer sessions, that's a legitimate choice; if splitting it across more frequent, shorter sessions fits your schedule (or your recovery) better, that works just as well, provided the total volume adds up to the same place. Proximity to Failure: Closer Helps a Little, All-Out Isn't Required This is the one where gym intuition and the research diverge most. According to PubMed, a systematic review and meta-analysis of 15 studies comparing training to different proximities to failure found only a trivial advantage for training to some form of "set failure" versus stopping short of it, and — more strikingly — found no advantage whatsoever for training all the way to momentary muscular failure (the true point of being unable to complete another rep) compared to stopping close to it ([Refalo et al., Sports Medicine, 2023, PMID: 36334240](https://doi.org/10.1007/s40279-022-01784-y)). The same analysis found no clear benefit to using higher velocity-loss thresholds (a more precise, technology-assisted way of measuring proximity to failure) over moderate ones. The authors describe this as evidence of a likely non-linear relationship: getting reasonably close to failure matters, but the last one or two reps of genuinely grinding to true failure don't appear to add proportional extra growth, despite how much additional fatigue and soreness they generate. This matters practically because training to true failure carries real costs — it increases fatigue, extends recovery time, and (particularly on compound lifts) raises injury risk, all for an effect the meta-analysis found statistically indistinguishable from stopping a rep or two short. Putting the Three Together None of these three findings argues for training carelessly — mechanical tension, sufficient weekly volume, and genuine effort are all still doing the real work. What the meta-analyses collectively push back on is the idea that there's one precise, narrow "correct" way to execute each of these variables. A wide range of rep tempos works. Weekly frequency is a flexible scheduling choice as long as volume holds constant. And stopping a rep or two shy of true failure appears to capture nearly all of the hypertrophic benefit without the compounding fatigue cost of grinding out every single set to its absolute limit. The Bottom Line If you're optimizing a training program based on gym folklore about perfect tempo, ideal frequency, or the necessity of failure training, the actual meta-analyses suggest most of that precision is solving a problem that doesn't meaningfully exist. Pick a rep speed that lets you control the weight, spread your weekly volume across however many sessions fits your schedule and recovery, and leave most sets a rep or two short of true failure. According to the current evidence, that combination captures the vast majority of the growth available, without the added fatigue of chasing precision on variables that turn out to matter far less than assumed.