Ask someone how hard a set was and they'll answer instantly and confidently. Ask them where that answer came from and the intuition is always the same: the muscle was working hard, so it told me it was working hard, and I felt that.
That intuition is wrong, and it's wrong in an interesting way.
Effort is a copy of the outgoing command
When your brain sends a motor command to a muscle, it also generates an internal copy of that command — a corollary discharge — which stays in the brain and is available to conscious perception. The sense of effort appears to be built primarily from that copy. You are not feeling the muscle. You are feeling how loudly you had to shout at it.
The cleanest demonstration of this comes from experiments where the muscle is artificially weakened — for instance by partially blocking the neuromuscular junction — while participants are asked to produce a given absolute force. The muscle is doing the same mechanical work it always did. The metabolic disturbance is not the driver. But because far more neural drive is now required to get that force out, the effort required feels dramatically greater. Weaken the output, and the same task feels harder despite nothing changing in the tissue.
It's worth separating three things that get bundled together and shouldn't be:
- Perceived effort — how hard you're driving. Central, from the motor command.
- Discomfort or burn — the metabolic sensation, signalled by group III and IV afferents responding to accumulated metabolites. Genuinely peripheral, genuinely from the muscle, and genuinely a different thing.
- Fatigue — the actual loss of force-producing capacity.
These correlate loosely enough to feel like one sensation and diverge often enough to ruin your training decisions. A high-rep set of leg extensions delivers enormous discomfort at moderate effort. A heavy triple delivers enormous effort with almost no burn. If you're using "how much did that hurt" as your proxy for "how hard did I train", you will systematically overrate the first and underrate the second.
Why the same weight isn't the same weight
Once you accept that effort tracks required neural drive rather than external load, a lot of previously annoying observations become predictions.
Anything that reduces your capacity to produce force means more drive is needed for the same bar weight, which means it feels harder. Poor sleep. An accumulated week of training. Being mid-deficit and glycogen-depleted. Being unwell. Having done four sets already. In every one of those cases the load on the bar is identical and the effort is not, and you are not imagining it — the perception is accurately reporting a real increase in required command.
It also runs the other way. Caffeine, high arousal, competition, music you actually like, a good night's sleep, being early in a session — all reduce perceived effort at a fixed load. Some of that comes with a genuine performance benefit and some of it is purely perceptual, which is precisely why "I felt strong today" is such an unreliable training metric on its own.
What RPE and RIR are actually for
Rating scales get treated as either gospel or nonsense, and they're neither. Their value is real but narrow.
The classical Borg scale was built for endurance work, where perceived exertion tracks cardiovascular and metabolic strain reasonably well. Applied to lifting it's a blunt instrument. The more useful version in resistance training is RIR-based RPE: rather than rating abstract difficulty, you estimate how many reps you had left. An 8 means two more were available. It reframes the question from "how did that feel" to "how close were you to the end", which is the thing that actually matters for stimulus.
The evidence on how accurately people do this is mixed in a fairly consistent pattern. Estimates get better as you get closer to failure — nobody can tell the difference between five and seven reps in reserve, but most people can tell one from three. They get better with training experience. They get worse at higher rep ranges, on unfamiliar exercises, and on isolation work where discomfort arrives early and impersonates failure.
So RIR is a decent instrument in the last few reps of a set and close to useless before that — which, conveniently, is the only region where you needed it.
Calibrating a gauge you can't read directly
Rate the set after it ends, not during it. Rating mid-set means rating your emotional response to how hard it currently is. Rating immediately after lets you assess what actually happened, and it's noticeably more accurate.
Anchor to real failure periodically. An RIR estimate is a comparison against a remembered reference, and references drift — usually toward comfort. Occasional genuine failure sets on safe exercises reset the reference. Without them, your 8s become 6s over the course of a block without a single conscious decision.
Distrust it most where the burn is worst. If an exercise makes you want to stop for metabolic reasons rather than mechanical ones, assume your RIR estimate is optimistic by at least two reps and act accordingly.
Cross-check against something external. Bar speed, reps at a given load over time, or simply whether the logbook is moving. If perceived effort is high and the objective numbers are flat for weeks, the effort rating is measuring your tolerance for discomfort rather than your proximity to a growth stimulus.
Don't chase the rating. The number is a description of a set, not a target to hit. Sets exist to produce a stimulus; RPE exists to help you notice whether one happened.
The wider point
Perceived effort isn't just a training-log field. It's one of the strongest predictors of whether a person continues exercising at all. How hard something feels — not how hard it is, not how effective it is — is a large part of what determines whether someone comes back next week. That's not a side note; that's most of the adherence problem, and it's where the psychology section picks the thread up.
The instrument is biased. It's also the only one you've got in real time, so the job isn't to discard it — it's to know the direction of its error and correct for it deliberately.