I was at the gym the other day when I watched a trainer put a client through a set. The client was grunting, then groaning, then outright screaming — veins out, face contorted, the whole performance. The trainer nodded along like this was the point.
Here's the thing that struck me: that wasn't effort 7. That was effort 11. And the research is fairly clear that for most people, most of the time, effort 11 isn't where the results live. It's where the fatigue lives.
Let me explain what I mean, because there's a number behind this, and it's one of the most useful things you can learn about training.
Effort has a scale, and most people misread it
Sports scientists measure how hard a set feels using two related ideas:
- RPE — Rating of Perceived Exertion, a 1–10 scale of how hard a set felt.
- RIR — Reps in Reserve, or how many more reps you couldhave done before your form broke down and you physically couldn't complete another.
They map onto each other. If you finish a set and could have done 2 more good reps, that's 2 RIR — which lands around RPE 8. Zero reps left in the tank, grinding to a dead stop? That's 0 RIR, RPE 10 — true failure. The screaming.
Most people think they're training at a 7 or 8. Watch them, and a lot of them are either coasting at a 4 (way too easy) or redlining at a 10 every set (way too hard). Both are mistakes. And the second one — the screaming — is the one dressed up to look like dedication.
What the research actually says
For years, gym lore held that you had to train to absolute failure — total, screaming, can't-move-the-bar failure — to build muscle. The evidence over the last few years has quietly dismantled that.
A 2024 study following trained lifters through eight weeks found that stopping a couple of reps short of failure produced the same muscle growth as grinding all the way to failure (Refalo et al., Journal of Sports Sciences, 2024). Same result, far less suffering.
The most comprehensive look at this came from a 2024 series of meta-regressions in the journal Sports Medicine, which pooled data from 55 muscle-growth studies and 67 strength studies and modeled effort as a sliding scale rather than a simple failure-or-not question (Robinson et al., Sports Medicine, 2024). Two findings matter for the screaming guy:
- For muscle growth, getting closer to failure helps — but the benefit flattens out past about 2 reps in reserve. In other words, the difference between stopping at 2 RIR and grinding to 0 is small and shrinking, while the fatigue cost keeps climbing.
- For strength, training closer to failure didn't reliably help at all — strength improved across a wide range of efforts. Some evidence even suggests grinding to failure can hurt strength by burying you in fatigue you then have to recover from (Robinson et al., 2024).
Put simply: the last, screaming, everything-you've-got rep is mostly buying you fatigue, not results.
So where should the effort actually sit?
For the large majority of your working sets, the productive zone is roughly RPE 7 to 9 — meaning you stop with about 1 to 3 good reps still in the tank. Hard enough that the last couple of reps are genuinely challenging and your form is starting to require real focus. Notso hard that you're screaming, your technique is falling apart, and you need five minutes and a spotter to recover.
That zone gives you nearly all of the muscle-growth stimulus, most of the strength benefit, and a fraction of the fatigue — which means you can train more often, recover better, and actually keep your form intact rep to rep. The screaming set might feel more heroic, but it's a worse trade almost every time.
There are narrow exceptions. An advanced lifter deliberately pushing a final set to failure on an isolation exercise, or someone training for muscular endurance, has reasons to occasionally touch that ceiling. But those are specific, planned choices — not the default, and not what should be happening on a general client's every set.
Why this matters more if you're over 40, injured, or on a GLP-1
Here's where it gets personal, and where the screaming approach goes from "suboptimal" to "actively counterproductive."
If you're over 40, managing an injury, coming back from surgery, or on a GLP-1 medication, your recovery capacity and your margin for error are different than a 22-year-old's. Grinding every set to screaming failure doesn't make you tougher — it digs a recovery hole you may not climb out of before your next session, raises your injury risk as form degrades on those last ugly reps, and makes the whole thing miserable enough that you quit. The research is clear you don't needthat intensity to grow muscle or get stronger. So for exactly the people the fitness industry tends to over-push, the smarter, sub-failure approach isn't just as good — it's genuinely better, because it's sustainable.
That's the whole philosophy behind how I think about coaching: the goal isn't to prove how much you can suffer in one session. It's to apply enough effort — that RPE 7–9 zone — consistently, for years, without breaking yourself. Effort you can repeat beats effort you have to recover from.
The screaming guy will have a great story about today's workout. The person training at a controlled 8 will have a better body in a year, and won't have hurt themselves getting there.